RELEASE NOTES FOR VERSION 2.8a

* INCOMPATIBLE CHANGE - The "language" option has new values. "language=autobib" changes
  language for bibliography entries based on the LANGID field and the "autolang" option setting.
  As a new feature, "language=autocite" does the same for citations. "language=auto" sets
  both "autocite" and "autobib" and the default is "lanauge=autobib". This is a change from
  the previous default. If your style sets "language=auto" explicitly, you will get the new
  citation language switching behaviour and might want to use "language=autobib" to get the old
  behaviour back.

RELEASE NOTES FOR VERSION 2.8

* Biber version 1.8 is required for biblatex 2.8
* Polyglossia is now better supported
* The HYPHENATION field is now called LANGID. The old name still works for backwards compat. This
  field does a lot more than just selecting hyphenation patterns and was misleading.
* New field LANGIDOPTS for Polyglossia users allows specification of
  language-specific options (like variants such as "american" english and
  babelshorthands etc.). See the biblatex-examples.bib file which has been
  converted to use the new field names and fields.
* "babel" option renamed to "autolang". Old name generates a warning but
  still works. Since Polyglossia support now basically works, this name is now
  too package specific. There is a new value "langname" for this option
  which is Polyglossia only and allows the use of the main polyglossia language
  switching environment which can use the options given in the LANGIDOPTS field.
* New value "year" for "datelabel" option
* New internal field "datelabelsource" for package authors specifies which
  date field the datelabel was generated from.

RELEASE NOTES FOR VERSION 2.7a

* Bugfix release for 2.7

RELEASE NOTES FOR VERSION 2.7

* Biber version 1.7 is required for biblatex 2.7

* New field EVENTTITLEADDON now part of default data model and default styles

* Many citation command enhancements, better flow-of-text integration for
  \textcite et al. See changelog in PDF doc for details.

RELEASE NOTES FOR VERSION 2.6

* Biber version 1.6 is required for biblatex 2.6

* INCOMPATIBLE CHANGE - \DeclareLabelyear is now \DeclareLabeldate so
  that the extra fields labelmonth and labelday can be generated by
  biber. The 'labelyear' package option is now called 'labeldate'. The
  old names will work but will generate warnings. The default
  definition for \DeclareLabeldate in biblatex.def results in the same
  behaviour as with the old \DeclareLabelyear.

* Localised and fixed string fallbacks now possible as values for labelyear field

* Cyclic and cascading RELATED entries now working properly.

RELEASE NOTES FOR VERSION 2.5

* Biber version 1.5 is required for biblatex 2.5

* New  sourcemap step "entrynull" can be used to completely skip an entry.

RELEASE NOTES FOR VERSION 2.4

* Biber version 1.4 is required for biblatex 2.4

* New macro \DeclareStyleSourcemap for style authors who want to define source mappings which
  are not changed by users using \DeclareSourcemap. Style maps so defined are run after user
  mappings and before driver default mappings.

* New RELATEDOPTIONS field giving more control over the exact options for related entry processing

* INCOMPATIBLE CHANGE - \DeclareLabelname, \DeclareLabelTitle and \DeclareLabelyear now have
  a different syntax in order to prepare for future multi-script support. The change is minor -
  just wrap all fields in \field{} and don't use any commas to separate fields. For example, the
  default definition of \DeclareLabelname was:

  \DeclareLabelname{%
    shortauthor,
    author,
    shorteditor,
    editor,
    translator}

  and is now:

  \DeclareLabelname{%
    \field{shortauthor}
    \field{author}
    \field{shorteditor}
    \field{editor}
    \field{translator}
  }

RELEASE NOTES FOR VERSION 2.3

*  Biber version 1.3 is required for biblatex 2.3

*  New "append" mode for \DeclareSourcemap so that fields can be combined

*  Detection of some situations which requires biber or latex reruns.
   Specifically, when sorting schemes are added or removed, when citations are added
   or removed, when sorting=none and citation order changes and the interactions of
   these situations when defernumbers=true. Biblatex should now report that a re-run is
   required in such situations.

*  Bugfixes

RELEASE NOTES FOR VERSION 2.2

*  Biber version 1.2 is required for biblatex 2.2

*  Bugfixes

RELEASE NOTES FOR VERSION 2.1

*  Biber version 1.1 is required for biblatex 2.1

*  Custom data models is now looked for in more places. See PDF
   documentation. Custom data model conflicts are now better processed by
   biber.

*  Some improved localisation (Norwegian) and doc fixes.

RELEASE NOTES FOR VERSION 2.0

*  Biber is now the default backend. Biblatex 2.0 requires biber
   1.0.  It will not work with any earlier release of biber due to
   .bcf format changes required to support the new list-local sorting
   feature.

*  Biblatex with bibtex as the backend is now frozen at version 1.7. When
   using biblatex with the option "backend=bibtex" (which must be
   specified now, the default is backend=biber if no option is specified),
   biblatex 1.7 will be automatically used. Apart from important bug fixes,
   no more development will be done on the 1.7 branch which means that
   users must switch to backend=biber (or don't specify the backend at all)
   to get version 2.0 and future new features. All major new 2.0 features
   require biber anyway so this is no real change for bibtex users.

*  New options to enable counters which track different combinations of
   information. Here is a summary which contains all tracking options,
   including the new ones (row marked with with a star, table is also in
   the PDF documentation):

                   Enabled     Enabled          Tracked
   Option          field       counter          information
   --------------  ----------  --------------   -------------------
   labelalpha      labelalpha  extraalpha       label
   labelyear       labelyear   extrayear        labelname+labelyear
   *labeltitle     ---         extratitle       labelname+labeltitle
   *labeltitleyear ---         extratitleyear   labeltitle+labelyear

   The naming of the options and counters is a little unintuitive because
   the original ones were named before requests for the new ones were made.
   The "Tracked information" column makes clear which information is tracked
   by the counters. The new counters will appear in the .bbl inside an entry
   and can be used just like "extraalpha" and "extrayear". The new options
   controlling the new counters are false by default and are not used by the
   standard styles.

*  The biblatex data model is now customisable using the \DeclareDatamodel*
   macros. This allows the use of new entrytypes and fields, as long as the
   style supports them.

*  Controlling initials generation

   A biblatex macro interface (\DeclareNoinit) is now available to declare
   regular expressions to strip from names before generating initials for
   them. This mirrors the Biber config file functionality on a per-document
   basis. A common use, for example, might be to strip certain prefices
   from names so that they do not end up as initials.

*  Fine tuning sorting strings

   A biblatex macro interface (\DeclareNosort) is now available to declare
   regular expressions to strip from fields or types of fields for sorting
   purposes. This mirrors the Biber config file functionality on a per-document
   basis. A common use, for example, might be to strip "The " from titles for
   sorting purposes.

*  Sorting option for bibliography/shorthand lists

   \printbibliography and \printshorthands now have a "sorting" option so
   you can choose a sorting scheme for each bibliography/shorthand list
   instead of using the global sorting scheme. This means that you can
   issue \printbibliography or \printshorthands as many times as you need
   to, anywhere in the document and have the resulting list sorted in the
   way required.

*  Dynamic datasource modification

   A biblatex macro interface (\DeclareSourcemap) is now available to
   allow users to dynamically modify data as it is read by Biber. This
   can be used to do many things like remove unwanted fields,
   regularise field contents, add new fields etc. all without write
   access to the datasource. The biblatex manual has a large section
   on this feature, with many examples. This feature is identical in
   functionality to Biber's "sourcemap" option but having a macro
   interface allows it to be used on a per-document basis. The default
   mappings for datasource drivers are definable too, using 
   \DeclareDefaultSourcemap
   
*  Customisable labels

   A new macro interface is implemented (\DeclareLabelalphaTemplate)
   which allows customisation of the label used in alphanumeric
   styles. The previous static labels are now implemented in terms of
   the new interface. The biblatex manual has a section on this
   feature with examples. One edge-case has changed with the new
   definition of the previous default label generation behaviour: a
   name with a prefix will now include 3 characters after the prefix
   instead of 2 ("van Rompel" will be "vRom" instead of "vRo"). The
   previous behaviour was anomolous anyway. There are options to
   auto-disambiguate labels in several ways.

*  Related entries

   A general method to support "related entries" is implemented. This
   allows a localisable, flexible way to deal with relationships between
   entries like "reprint of", "translation of", "reprinted in" etc.

   Three new entry fields are available to support this - "related",
   "relatedtype" and "relatedstring". Biber will automatically create
   "dataonly" clones of the related entries so that their data can be
   accessed in styles without having to cite them. Related entries may
   also be cited themselves which does not interfere in any way with this
   mechanism. The standard styles have been updated to support the new
   fields. New punctuation macros are available to format the new fields
   in the bibliography ("relatedpunct", "relateddelim", "related").

*  Citation key aliases

   It is now possible to include a new field "ids" in an entry which
   is a comma-separated list of citation key aliases. The entry can be
   cited by any of the aliases, which helps when you updated your
   entry keys but don't want to change old documents.

*  Option to sort names only using initials

   The new option "sortfirstinits" allows users to specify that name
   sorting only uses initials instead of the full name. Previously,
   even if "firstinits" was true, sorting used full names.

*  New citation commands

   Starred variants of \citeyear and \citedate are now available which
   include the extrayear information.

RELEASE NOTES FOR VERSION 1.7

*  Biber update, new features

   Biblatex 1.7 requires Biber 0.9.6. Biber is now based on Perl 5.14
   with full Unicode 6.0 support. There's a bunch of other
   improvements -- see the Biber manual for details.

*  XDATA containers [Biber only]

   This release introduces the new @xdata entry type and the
   corresponding xdata field. Essentially, the idea is that you can
   dump frequently used data into an @xdata entry and reference that
   where needed. It's a bit like crossref but doesn't imply any
   logical (parent/child) relation. Or, to put it differently, it's a
   bit like @string but covers more than one field. See the changelog
   for pointers.

*  Extended option scoping [Biber only]

   The package options related to name/list truncation and name
   disambiguation are now conveniently settable globally/per-type/
   per-entry. See the changelog for details. There's also a new
   appendix with an overview of the scope of all options.

*  Field formats for dates

   \printdate and friends now support 'virtual' field formats for all
   dates. So, even though there is no 'date' field on the LaTeX end
   of the workflow (only year/month/day and endyear/endmonth/endday),
   you may now say:

   \DeclareFieldFormat{date}{\textbf{#1}}
   
   to print the date in boldface, for example. Technically, this is
   equivalent to:

   \DeclareFieldFormat{date}{\textbf{#1}}
   \printtext[date]{\printdate}
   
   except that it is readily built into \printdate.

*  Last-resort breakpoints for \url

   URLs and DOIs consisting mostly or entirely of letters and numbers
   may run into the margin if the \url command can't find suitable
   breakpoints. I've introduced additional breakpoints after all
   numbers and letters in the previous biblatex release in order to
   deal with particularly nasty DOIS, but this turned out to be to
   brute a solution.

   To provide the best of both worlds, this release adds counters
   which control the penalties associated with the additional
   breakpoints. They are disabled by default but may be helpful (as a
   last resort) if you need to deal with long and/or tricky DOIs. See
   'biburlnumpenalty' etc. in the manual for details. If you want to
   enable the additional breakpoints, use high penalties; e.g.:
   
   \setcounter{biburlnumpenalty}{9000}

*  Language support

   Oleg Domanov contributed support for Russian. Note that the
   Russian module requires UTF-8. Pieter Belmans updated the Dutch
   language module.

RELEASE NOTES FOR VERSION 1.6

*  Biber update, new mapping feature

   Biblatex 1.6 requires Biber 0.9.4. It will not work with any
   previous release. Biber now supports user-space field mapping.
   This means that you can set up mapping rules in the Biber
   configuration file to deal with non-standard field names in .bib
   files by mapping them to known field names.

*  Revised maxnames/minnames options

   Biber now supports max/mincitenames and max/minbibnames as
   separate values. This improvement has prompted a refactoring of
   the corresponding package options. In previous versions, there
   were three sets of values:

   max/minnames     -> backend+frontend default
   max/mincitenames -> frontend only
   max/minbibnames  -> frontend only

   The backend would get max/minnames, the other values were only
   used on the LaTeX end, defaulting to max/minnames. Depending on
   the citation style, this could lead to ambiguous citations if
   max/minnames was out of sync with max/mincitenames. (This wasn't
   really obvious from reading the manual.)

   This can no longer happen. The new solution is simpler and
   error-proof in this respect. The max/mincitenames and
   max/minbibnames options do what their name indicates and are fully
   supported by Biber. The max/minnames options now serve as master
   options which set both max/minbibnames and max/mincitenames:

   max/minnames     -> max/mincitenames
                    -> max/minbibnames
   max/mincitenames -> frontend+backend
   max/minbibnames  -> frontend+backend

   We didn't bother to update the BibTeX backend as it is due to be
   discontinued this fall anyway. The BibTeX backend will get the
   max/mincitenames values only.

*  [CAVEAT] Removed local max/minnames and max/minitems options

   As part of the above change, I've removed the local max/minnames
   and max/minitems options from \printbibliography and friends to
   enforce consistency. Please use the global options instead.

*  Added maxalphanames/minalphanames [Biber only]

   With Biber, the 'labelalpha' field may now be configured
   independently of max/mincitenames with the newly introduced
   package options max/minalphanames.

*  MD5 hashes, per-name hashes [Biber only]

   With Biber, the hashes provided in the namehash/fullhash fields
   are now standard MD5 hashes. This means that they are stable
   across multiple runs, multiple refsections, and multiple
   documents.

   We've also added per-name hashes. Every name in every name list
   gets an MD5 hash, which is available in the 'hash' field. Note
   that this field is only available locally in name formatting
   directives.

*  Range compression/truncation

   This release features a new command for range compression (i.e.,
   format "123-128" as "123-8"). The command is configurable, can
   handle lists of ranges separated by commas and/or semicolons, and
   also normalizes the dashes. Note that combining this with
   \mkpageprefix requires the field format to be set up in a specific
   way. See \mkcomprange in the manual for details.

   I've also overhauled the command which truncates ranges by only
   printing the starting page of the range. The original 
   \mkpagefirst has been removed because of its awkward syntax. It is
   superseded by \mkfirstpage. Note that this change is not
   backwards-compatible because the syntax of the new command is
   slightly different. You may need to update some format
   definitions.

RELEASE NOTES FOR VERSION 1.5

*  Biber update

   Biblatex 1.5 requires Biber 0.9.3. It will not work with any
   previous release.

*  More name disambiguation modes [Biber only]

   We've added some additional modes to the 'uniquename' and
   'uniquelist' options. Roughly speaking, the new 'min*' modes aim
   for unambiguous citations (mainly author-year) rather than
   disambiguating names in a strict way. There's a section in the
   manual which discusses all the details. See the changelog for
   pointers to the relevant parts of the manual.

RELEASE NOTES FOR VERSION 1.4

*  Biber update

   Biblatex 1.4 requires Biber 0.9 or higher. It will not work with
   any previous release.

*  Advanced name disambiguation in citations [Biber only]

   Name disambiguation in citations has been greatly improved and
   extended in this release. We're pretty excited about this feature
   because it has been in the pipeline for such a long time and
   because Biblatex/Biber is, to the best of my knowledge, the only
   bibliographic tool providing fully automated disambiguation of
   names in citations.

   There are now two options controlling name disambiguation,
   'uniquename' and 'uniquelist'. You can use either option
   stand-alone or combine both. This is best explained by way of
   example. Consider the following data:

   John Smith/John Doe   2011
   John Smith/Jane Doe   2011
   John Smith/Edward Doe 2011
   
   Let's assume we're using the standard 'authoryear' style and set
   uniquename=false. In this case, we would get the following
   citations:

   Smith and Doe 2011a
   Smith and Doe 2011b
   Smith and Doe 2011c

   The extra letter appended to the year (that's the automatically
   generated 'extrayear' field provided by the 'labelyear' option)
   disambiguates the citations. Many style guides, however, mandate
   that the extra letter be used to disambiguate works by the same
   author(s) only, not works by different authors with the same last
   name. In order to disambiguate the author's last name, you're
   expected to add additional parts of the name, either as initials
   or in full. This is what the 'uniquename' option does. Here are
   the same citations with uniquename=init:

   Smith and J. Doe 2011a
   Smith and J. Doe 2011b
   Smith and E. Doe 2011
   
   'uniquename=init' restricts name disambiguation to initials. Since
   the first two Does are still ambiguous at that point, we need to
   fall back to 'extrayear'. With uniquename=full, names are printed
   in full where required:

   Smith and John Doe 2011
   Smith and Jane Doe 2011
   Smith and E. Doe 2011
   
   In the past, this feature was of limited use since it only worked
   for entries with a single author/editor. It now works for any name
   at any point in the list. Since this is only possible with Biber,
   the feature has been removed entirely from the BibTeX backend.

   When it comes to lists of names, there's another aspect to
   ambiguity. Let's call it list ambiguity. Consider the following
   data:

   Doe/Jones/Smith    2011
   Smith/Johnson/Doe  2011
   Smith/Doe/Edwards  2011
   Smith/Doe/Jones    2011
   
   Most author-year and author-title styles will truncate long
   author/editor lists in citations. For example, with maxnames=1
   we'd get:

   Doe et al. 2011
   Smith et al. 2011a
   Smith et al. 2011b
   Smith et al. 2011c
   
   Since the author lists are ambiguous after truncation, the
   'extrayear' marker is added by the style in order to ensure unique
   citations. Here again, many style guides mandate that the extra
   letter be used to disambiguate works by the same authors only. In
   order to disambiguate author lists, you're typically required to
   add more names, exceeding the maxnames/minnames truncation point.
   The new 'uniquelist' feature introduced in this release does that
   automatically. With maxnames=1 and uniquelist=true, we get:

   Doe et al. 2011
   Smith, Johnson et al. 2011
   Smith, Doe and Edwards 2011
   Smith, Doe and Jones 2011
   
   As seen above, the 'uniquelist' option overrides the 'maxnames'
   setting on a per-entry basis. Essentially, what happens is that
   the "et al." parts of the citations get expanded to the point of
   no ambiguity -- but no further than that.

   'uniquename' and 'uniquelist' may also be combined. Consider the
   following data:

   John Doe/Jane Smith/William Jones    2011
   John Doe/John Smith/William Jones    2011
   John Doe/Edward Smith/William Jones  2011
   John Doe/John Edwards/William Jones  2011
   John Doe/John Edwards/Jack Johnson   2011

   With maxnames=1:
   
   Doe et al. 2011a
   Doe et al. 2011b
   Doe et al. 2011c
   Doe et al. 2011d
   Doe et al. 2011e

   With maxnames=1, uniquename=true, uniquelist=true:
   
   Doe, Jane Smith et al. 2011
   Doe, John Smith et al. 2011
   Doe, E. Smith et al. 2011
   Doe, Edwards and Jones 2011
   Doe, Edwards and Johnson 2011

   If multiple disambiguation options are active, names are
   disambiguated first and works last. The order is as follows:

   1. uniquename
   2. uniquelist
   3. extrayear, singletitle

   Adding this feature to existing styles is simple. The 'uniquename'
   information needs to be evaluated in the name formatting directive
   which controls the formatting of the 'labelname' list. The default
   definition in biblatex.def is responsive to the 'uniquename'
   option. If you're using that, just enable the package option and
   that's it. If you're using a different format, check biblatex.def
   on how to incorporate the 'uniquename' feature.
   
   Adding support for 'uniquelist' is also trivial. If the second
   optional argument of \printnames is omitted (i.e., no explicit
   range is specified), \printnames{labelname} will print the
   'labelname' list up to the truncation point defined with the
   'maxnames'/'minnames' package options. If 'uniquelist' is enabled,
   it will simply override 'maxnames' on a per-entry basis, i.e.,
   \printnames will print the 'labelname' list up to point of no
   ambiguity. Unless you're specifying explicit ranges, all you need
   to do is enable the 'uniquelist' package option.

*  Name disambiguation in standard styles [Biber only]

   All standard author-year and author-title styles now enable
   uniquename/uniquelist disambiguation. Of course this only works
   when using Biber.

RELEASE NOTES FOR VERSION 1.3

The list of changes in Biblatex is fairly short this time. Most of
the big feature improvements are found in Biber.

*  Biber 0.8.3, Biber in TeX Live main.

   Update Biber to version 0.8.3 when upgrading to biblatex 1.3. Note
   that Biber is now available from the main TeX Live repository.
   This means that you don't need to point tlmgr to TLContrib any
   more. It also implies that more architectures are supported.

*  Experimental Zotero and EndNote import [Biber only]

   Biber now features import support for the Zotero RDF/XML and
   EndNote XML formats. See the Biber manual for details and caveats.

*  Fine-grained name delimiter/spacing controls [Biber only]

   The spacing between name parts and line breaks in names can now be
   customized in great detail, especially with Biber. With BibTeX,
   customization is much more limited but still possible to some
   extent. See the manual for details.

*  Added citation counter

   As its name suggests, the new 'citecounter' feature counts
   citations, i.e., it indicates how many times an entry is cited in
   a reference section. The value of the 'citecounter' counter refers
   to the total per-entry/per-refsection citation count.

   The counter works much like the citation trackers. It is useful
   for styles which handle initial and repeated citations
   differently. For example, you may want to suppress the
   introduction of a shorthand of an entry in the initial citation if
   it is cited only once anyway.

*  Added \smartcite command

   I've added a new citation command called \smartcite (along with
   \Smartcite, \smartcites, etc.). Essentially, this is a \footcite
   command which behaves like \parencite when used in a footnote.
   It's particularly useful in conjunction with autocite=footnote as
   it makes \autocite even more versatile.

   Style authors should note that the default setup for
   autocite=footnote has been modified and now uses \smartcite
   instead of \footcite. Providing a \smartcite command is
   straightforward. In most cases, all you need to do in your style
   is duplicate the definition of \footcite and adapt the wrapper
   code like so:

   \DeclareCiteCommand{\smartcite}
     [\iffootnote{\mkbibparens}{\mkbibfootnote}]
     {...}
     {...}
     {...}
     {...}

RELEASE NOTES FOR VERSION 1.2

*  Remote resources, experimental RIS import [Biber only]

   Biber 0.8 features support for remote .bib files (HTTP/FTP) and
   experimental RIS file import. See the Biber manual and
   \addbibresource in the biblatex manual for details.

*  \bibliography superseded by \addbibresource

   As Biber development opens new possibilities, I've refactored the
   user interface for specifying bibliographic resources.
   \bibliography is superseded by \addbibresource. There are two
   differences when using \addbibresource:

   1) Always give the full file name, do not omit the .bib extension.
   2) No comma-separated lists, use one \addbibresource per bib file.
   
   For example, the old form:
   
   \bibliography{file1,file2}
   
   is superseded by:
   
   \addbibresource{file1.bib}
   \addbibresource{file2.bib}

   \addbibresource also supports options:

   \addbibresource[location=local,type=file,datatype=bibtex]{file1.bib}

   You only need to specify options which differ from the defaults
   (local/file/bibtex), e.g.:

   \addbibresource[datatype=ris]{file.ris}
   \addbibresource[location=remote]{ftp://192.168.3.57/~fred/file1.bib}

   The change is fully backwards compatible since \bibliography is
   still available. There is no need to update existing files but
   \addbibresource is the recommended way for new files.

*  Smart/customizable 'crossref' data inheritance [Biber only]

   Biber has been supporting 'crossref' customization for some time
   but the biblatex user interface was still missing. This release
   finally adds a user interface. This means that you may configure
   how the data is inherited from the entry the 'crossref' field
   refers to. The manual has all the details, including the default
   settings.

   If you're using the 'crossref' field, this change is a significant
   one since the smart data inheritance is set up by default. By
   'smart' I mean a sensible parent/child field mapping which
   replaces BibTeX's symmetric mapping. There is no need any more to
   duplicate fields.

   BibTeX crossref:

   book		->	inbook
   -----------------------------------
   title	->	title
   booktitle	->	booktitle

   Biber crossref:

   book		->	inbook
   -----------------------------------
   title	->	booktitle

*  New multi-volume entry types

   There are four new entry types in this release: mvbook,
   mvcollection, mvproceedings, mvreference. They are similar to the
   established types without the 'mv' prefix but indicate a
   multi-volume work explicitly.

   This difference is crucial of you want to leverage Biber's
   'crossref' support because titles are inherited differently in
   this case:

   book		->	inbook
   -----------------------------------
   title	->	booktitle
   subtitle	->	booksubtitle
   titleaddon	->	booktitleaddon

   mvbook	->	inbook
   -----------------------------------
   title	->	maintitle
   subtitle	->	mainsubtitle
   titleaddon	->	maintitleaddon
   
RELEASE NOTES FOR VERSION 1.1

Starting with this release, we'll leverage the possibilities of Biber
to support features not possible with BibTeX. That's why most major
new features in this release are 'Biber only'.

*  Biber update

   Update Biber to version 0.7 when installing Biblatex 1.1. The
   latest Biber is available from this location:

   http://sourceforge.net/projects/biblatex-biber/files/
     biblatex-biber/current/

   Ready-to-run binaries are here:
   
   http://sourceforge.net/projects/biblatex-biber/files/
     biblatex-biber/current/binaries/

*  Configurable sorting schemes [Biber only]

   This release introduces configurable sorting schemes. With Biber,
   you may now set up arbitrary sorting specs. The new custom sorting
   specs are very powerful and come with a flexible user interface.
   See \DeclareSortingScheme, \DeclareSortExclusion, and
   \DeclarePresort in the manual for details. Like the predefined
   sorting specs, new ones are activated with the 'sorting' option.

   If you are using Biber, all known sorting schemes are still
   available by default, but they are now defined in biblatex.def
   instead of being hard-coded in the backend.

   If you are using BibTeX, note that the old hard-coded schemes in
   biblatex.bst have not been altered. They continue to work as they
   did before, but you can't modify them or define new ones.

*  Configurable 'labelname'/'labelyear' [Biber only]

   The special fields 'labelname'/'labelyear' may now be customized
   by style authors. In conjunction with the configurable sorting
   schemes, this dramatically increases biblatex's flexibility,
   especially with author-year, author-title, and verbose styles. See
   \DeclareLabelname and \DeclareLabelyear in the manual for details.

*  Introducing per-type package options

   Starting with this release, certain package options may be set on
   a per-type basis. Use the new optional argument of
   \ExecuteBibliographyOptions to specify the type. Note that most
   per-type options are 'Biber only'. For details, see the rather
   lengthy list of per-type options in the changelog.

*  Improved static entry sets [Biber only]

   Static entry sets (i.e., @set entries in the bib file) are now
   natively supported by Biber. This means that instead of the rather
   convoluted old setup:
   
   @Set{set1,
     entryset = {key1,key2,key3},
     crossref = {key1},
   }
   @Article{key1,
     entryset = {set1},
     author   = {...},
     title    = {...},
     ...
   }
   @InCollection{key2,
     entryset = {set1},
     author   = {...},
     title    = {...},
     ...
   }
   @Article{key3,
     entryset = {set1},
     author   = {...},
     title    = {...},
     ...
   }

   defining a static set is now as simple as this:
   
   @Set{set1,
     entryset = {key1,key2,key3},
   }

   The 'crossref' in the @set head entry and the 'entryset' pointers
   in the member entries are no longer required. Defining a static
   set is now as simple as adding a @set entry with an 'entryset'
   field specifying the set. That's all.

   Since there are no 'entryset' pointers in the member entries, the
   members may be part of a set in one document/refsection and
   stand-alone references in another one, depending on the presence
   of the @set entry. If the @set entry is cited, the set members are
   grouped automatically. If not, they behave like regular entries.

*  Introducing dynamic entry sets [Biber only]

   Dynamic entry sets are a new way of grouping entries. They differ
   from static ones in that they are not defined in the bib file but
   declared in the document, i.e., they work on a per-document/
   per-refsection basis. Instead of defining a set in the bib file:

   @Set{set1,
     entryset = {key1,key2,key3},
   }
   
   you declare it like that in the document:

   \defbibentryset{set1}{key1,key2,key3}
   
   This works anywhere in the document, including the document body.
   Apart from that, dynamic entry sets work exactly like static ones.
   See the manual for more information.

*  mcite/mciteplus-like citation commands [Biber only]

   Users of the mcite/mciteplus packages are already familiar with
   the concept of a dynamic entry set. With mcite/mciteplus, sets are
   defined as they are cited. To facilitate migration to biblatex,
   this release comes with a special citation module which provides
   mcite/mciteplus-like citation commands on top of the commands
   provided by the main citation style. For example, the following
   command:

   \mcite{setA,*keyA1,*keyA2,*keyA3}

   is equivalent to this:

   \defbibentryset{setA}{keyA1,keyA2,keyA3}\cite{setA}
   
   Note that there is a syntactical difference between biblatex's
   dynamic entry sets and the citation grouping of mcite/mciteplus.
   With mcite/mciteplus, the first set member plays double-duty as
   group identifier. With biblatex, an entry set is an entity in its
   own right which requires a unique entry key.

   For example, an mcite citation group is declared like this:

   \cite{keyA1,*keyA2,*keyA3}
   
   In contrast to that, a dynamic entry set defined with one of the
   commands provided by biblatex's 'mcite' package option is set up
   like that:

   \mcite{setA,*keyA1,*keyA2,*keyA3}
   
   Note the extra identifier 'setA': this is the entry key assigned
   to the set. See the manual for further details.

*  Low-level bibliography filters

   The \defbibfilter command and the 'filter' option of
   \printbibliography are now supplemented by \defbibcheck and the
   corresponding 'check' option.
   
   Conceptually, a 'check' is like a filter controlling which entries
   are included in a certain (sub-)bibliography. In contrast to
   filters, however, the checks don't use a high-level syntax. Checks
   are macros which directly employ the low-level conditionals of
   biblatex's data interface. Here's an example 'check' which tests
   if an entry has an abstract:

   \defbibcheck{hasabstract}{%
     \iffieldundef{abstract}{\skipentry}{}%
   }
   ...
   \printbibliography[check=hasabstract]
   
   This will only print entries with an 'abstract' field. More complex
   checks are possible as well. For example, this check will exclude
   all entries published prior to the year 2000:

   \defbibcheck{recent}{%
     \iffieldint{year}
       {\ifnumless{\thefield{year}}{2000}
	  {\skipentry}
	  {}}
       {\skipentry}}

   See the manual for further details.

*  More flexible 'maxnames'/'minnames' controls

   I've added some additional package options related to 'maxnames'/
   'minnames'. See 'maxbibnames'/'minbibnames' and 'maxcitenames'/
   'mincitenames' in the manual.

   The new options don't introduce entirely new functionality. They
   are rather convenience options which allow setting 'maxnames'/
   'minnames' in a more flexible way. For example, setting the
   'maxbibnames' option like this:

   \usepackage[maxnames=3,maxbibnames=99]{biblatex}
   
   is equivalent to:
   
   \usepackage[maxnames=3]{biblatex}
   \begin{document}
   ...
   \printbibliography[maxnames=99]
   
   'maxcitenames' does the same for all citations.
   
*  Defining and resetting per-type formats

   \DeclareFieldFormat and related commands support per-type
   formatting directives. By default, biblatex uses this feature to
   format the 'title' field depending on the entry type. For example,
   book titles are printed in italics, article titles use quotes
   instead. In previous releases, lines like the following ones were
   found in biblatex.def:

   \DeclareFieldFormat{title}{\mkbibemph{#1}}
   \DeclareFieldFormat[article]{title}{\mkbibquote{#1}}
   \DeclareFieldFormat[inbook]{title}{\mkbibquote{#1}}
   ...

   This release brings two new features related to that.

   1) \DeclareFieldFormat and related commands now support a
   comma-separated list of entry types. This facilitates the
   definition of per-type formats:

   \DeclareFieldFormat{title}{\mkbibemph{#1}}
   \DeclareFieldFormat[article,inbook,...]{title}{\mkbibquote{#1}}

   2) \DeclareFieldFormat and friends now come with a starred variant
   which resets all type-specific formats. With previous releases,
   removing or modifying the formatting could be tedious if you
   didn't make that book/article distrinction. You had to change the
   default plus all type-specific formats:

   \DeclareFieldFormat{title}{#1}
   \DeclareFieldFormat[article]{title}{#1}
   \DeclareFieldFormat[inbook]{title}{#1}
   ...
   
   The new starred commands do that in one shot:

   \DeclareFieldFormat*{title}{#1}
   
   Note that the default formatting has not been altered. It's just a
   bit easier to modify it now.

*  New option 'mergedate' (authoryear style)

   By popular request, I've added a 'mergedate' option to the
   'authoryear' bibliography style. The option controls whether or
   not the date label is merged with the date specification in the
   bibliography.

   For example, while mergedate=false outputs:

   Doe, John (2008a). Article. In: Journal 23.5 (May 2008),
   pp. 5-24.

   mergedate=true will print:

   Doe, John (May 2008a). Article. In: Journal 23.5, pp. 15-34.

   Note that 'mergedate=true' is in fact the old behavior (and the
   default setting). 'mergedate=false' is the new feature!

RELEASE NOTES FOR VERSION 1.0

Here it finally is, the long-awaited first stable release of
biblatex. As the 0.9* releases have proven to be very mature and
reliable, this release is essentially 0.9e with some localization
updates and minor tweaks. It concludes the 0.9 development cycle
rather than introducing completely new features.

The stable status basically has two consequences. 1) We'll try to be
more cautious in terms of backwards compatibility in the future and
2) Biblatex moves to a different location on CTAN. See below for
details.

*  Development Roadmap

   Over the course of the 1.x development cycle, you'll increasingly
   see new features being added which are tagged as 'Biber only' in
   the manual. The BibTeX backend of Biblatex will be maintained for
   some more time but it won't receive any feature updates. There is
   a simple reason for that: the kind of features we'll introduce
   can't be handled by BibTeX (that's why we have Biber, after all).

   While you can stay with the BibTeX backend for some more time,
   it's highly advisable to switch to Biber fairly soon as BibTeX
   support will be discontinued at some point. To give you an idea
   of the timescale: we're talking months rather than years.

*  New location on CTAN

   The package moves from

	tex-archive/macros/latex/exptl/biblatex/
   to
	tex-archive/macros/latex/contrib/biblatex/
   
   The pre-bundled TDS archive moves from
   
	tex-archive/install/macros/latex/exptl/
   to
	tex-archive/install/macros/latex/contrib/

*  New contrib location on CTAN

   There's also a new location for contributed styles:
   
	tex-archive/macros/latex/contrib/biblatex-contrib/

   The old exptl/biblatex-contrib/ directory is still available and
   will keep being available as a staging area for styles which are
   work in progress. If your style is considered stable, you may
   want to move it to contrib/biblatex-contrib/. If it's
   experimental, you may prefer exptl/biblatex-contrib/.

*  Biber update

   When updating Biblatex, make sure you also get the latest Biber
   from this location:

   http://sourceforge.net/projects/biblatex-biber/files/
     biblatex-biber/current/

   Ready-to-run binaries are here:
   
   http://sourceforge.net/projects/biblatex-biber/files/
     biblatex-biber/current/binaries/

*  Modified 'bibencoding' package option, new default

   I've renamed 'bibencoding=inputenc' to 'bibencoding=auto' to make
   it clear that this option is not specific to the inputenc package
   and also works with XeTeX/LuaTeX in native UTF-8 mode.
   bibencoding=inputenc is still supported as an alias.

   'bibencoding=auto' is the new package default. This means that
   biblatex by default assumes that the workflow is transparent,
   i.e., that the .bib files use the same encoding as the .tex file.
   Note that it is the input encoding of the .tex (!) file which is
   detected automatically. If the .bib files use a different
   encoding, you always need to specify it explicitly.
   
   If you prefer the old default, you can easily restore it by
   adding the following like to your biblatex.cfg file:

   \ExecuteBibliographyOptions{bibencoding=ascii}

   The new default setting may trigger some warnings which have not
   been triggered before. Consider this case:

   \usepackage[latin1]{inputenc}
   \usepackage{biblatex}
   
   This will trigger a warning instructing you to switch to bibtex8
   or Biber. That's because the defaults are 'backend=bibtex' and
   'bibencoding=auto'. The latter implies that the .bib files are
   Latin1 encoded but traditional BibTeX can't handle that, hence
   the warning. If the .bib files do in fact use Latin1 encoding it's
   a good idea to heed the advice:
   
   \usepackage[latin1]{inputenc}
   \usepackage[backend=bibtex8/biber]{biblatex}
   
   If they're Ascii, set the encoding explicitly:

   \usepackage[latin1]{inputenc}
   \usepackage[bibencoding=ascii]{biblatex}
   
   On the other hand, you may delete all old 'bibencoding=inputenc'
   options as this is the default setting now.
   
RELEASE NOTES FOR VERSION 0.9e

*  Biber update, options 'sortupper', 'sortlocale'

   Biber has been updated to version 0.5.7. There are ready-to-run
   binaries for the latest version. If you're using Biber, upgrade
   to 0.5.7 or later when upgrading to biblatex 0.9e.

   Biblatex now offers two package options which correspond to
   Biber's command line options --sortupper and --sortlocale, the
   point being that you can set them on a per-document basis. I've
   also added some explanations concerning encodings and locale
   settings. See the changelog for pointers.

*  Improved backrefs for @set entries, 'backrefsetstyle' option

   I've revised and improved the handling of back references related
   to entry sets. Biblatex is now capable of tracking the @set head
   entry and the set members separately, if so desired.
   
   The new package option 'backrefsetstyle' controls the tracking
   mode. By way of example, consider a set ("set1") with three
   members ("memA", "memB", "memC") and the following input:

   on page 1: \cite{set1}
   on page 2: \cite{memA}
   on page 3: \cite{memB}
   on page 4: \cite{memC}

   With backrefsetstyle=setonly, this will generate the following back
   references:
   
   set1: 1, 2, 3, 4
   memA: -
   memB: -
   memC: -

   With backrefsetstyle=memonly:
   
   set1: -
   memA: 1, 2
   memB: 1, 3
   memC: 1, 4

   With backrefsetstyle=setormem:
   
   set1: 1
   memA: 2
   memB: 3
   memC: 4

   With backrefsetstyle=setandmem:
   
   set1: 1, 2, 3, 4
   memA: 2
   memB: 3
   memC: 4

   With backrefsetstyle=memandset:
   
   set1: 1
   memA: 1, 2
   memB: 1, 3
   memC: 1, 4

   With backrefsetstyle=setplusmem:
   
   set1: 1, 2, 3, 4
   memA: 1, 2
   memB: 1, 3
   memC: 1, 4
   
   To take advantage of that in custom styles, only a minor update
   is required. Supporting the @set entry and the set members
   separately is pretty intuitive. Any \printlist command in the
   @set entry will see the 'pageref' list of the @set entry; the
   \printlist commands in the regular drivers (which handle the set
   members) will see the 'pageref' list of the respective member:

   \DeclareBibliographyDriver{set}{%
     \entryset{}{}%
     \newunit\newblock  % <- NEWLY ADDED
     \printlist{pageref}% <- NEWLY ADDED
     \finentry}

   \DeclareBibliographyDriver{<entry type>}{%
     ...
     \printlist{pageref}%
     ...
   }
   
   There are some bibmacros as well as a new \bibpagerefpunct
   separator to help with that, hence the actual code in the
   standard styles works like this:

   \DeclareBibliographyDriver{set}{%
     \entryset{}{}%
     \newunit\newblock       % <- NEWLY ADDED
     \usebibmacro{setpageref}% <- NEWLY ADDED
     \finentry}

   \DeclareBibliographyDriver{<entry type>}{%
     ...
     \setunit{\bibpagerefpunct}\newblock % <- MODIFIED
     \usebibmacro{pageref}%
     ...
   }
   
   That's all. This will ensure that all tracking modes are
   supported.

RELEASE NOTES FOR VERSION 0.9d

*  Date formatting options

   I've removed all '<date>=none' options as this feature seems to
   be causing some confusion. Setting '<date>=none' was not
   sufficient to suppress a certain date in all cases because date
   specifications may also incorporate elements such as the 'issue'
   field, be tied to the 'url' in a way that leads to erroneous
   formatting if there is no date, etc. I'm afraid selectively
   omitting dates really needs to be handled on the style level.

   There's a new set of '<date>=iso8601' options which print
   extended-format ISO-8601 dates. Not all of ISO-8601 is
   implemented, only the core elements of the extended format which
   are also used in .bib files, i.e., "yyyy-mm-dd" for individual
   dates and "yyyy-mm-dd/yyyy-mm-dd" for ranges. Full-scale ISO-8601
   date/time specs like "2005-08-09T18:31:42P3Y6M4DT12H30M17S"
   wouldn't make much sense in a bibliography anyway. Note that
   these options are language agnostic. They will override any
   region-specific settings provided by .lbx files. That's the
   whole point of ISO-8601, after all. 

RELEASE NOTES FOR VERSION 0.9c

*  Improved reencoding support

   Biber 0.5.5 is capable of reencoding the data in bib files on the
   fly. This solution is more robust than biblatex's macro-level
   reencoding with inputenc. With 'backend=biber', biblatex will now
   use Biber instead of inputenc. In addition to that, the biblatex
   code which detects if the bib data needs to be reencoded to match
   the encoding of the tex file is smarter now and should work
   smoothly.

   In practice, this simply means that you may always use the
   'bibencoding' option to specify the encoding of your bib files
   (there was some potential for spurious reencoding steps in
   previous versions), with any backend. It is in fact highly
   recommended to do so for two reasons: 1) the default value is
   'ascii', and 2) biblatex will warn you if the bib encoding can't
   be handled by the backend you selected.

   You may put your platform default in biblatex.cfg; there is no
   need to do that in each and every preamble (you may still
   override all biblatex.cfg settings on a per-document basis in the
   preamble, if required).

*  UTF-8 encoded entry keys and bib filenames

   Biber 0.5.5 supports UTF-8 encoded entry keys. Biblatex will work
   with UTF-8 keys and filenames when running on an engine with
   native UTF-8 support (i.e., XeTeX or LuaTeX). Macro-level UTF-8
   support with inputenc/inputenx will not do.

*  Improved arXiv support

   I've added an additional eprint-related field plus two aliases to
   better support arXiv references. The changelog includes pointers
   to the relevant sections of the manual.

RELEASE NOTES FOR VERSION 0.9b

*  Extended Biber interface

   If you're using Biber as backend, make sure you update Biber as
   well (to version 0.5.4 or higher) because the Biblatex -> Biber
   interface has been extended.
   
   The most obvious effect of this improvement is that Biber is now
   capable of handling multiple refsections in a single pass. If you
   have any 'refsection' environments in a document, you only need to
   run Biber once.

*  New dependency: logreq package

   Starting with this release, biblatex will use the logreq package
   to write machine-readable messages to an auxiliary log file. When
   installing biblatex 0.9b, make sure to install the logreq package
   as well. It is available from CTAN:
   
   http://www.ctan.org/tex-archive/macros/latex/contrib/logreq/
   
   A TDS archive is available as well:
   
   http://www.ctan.org/tex-archive/install/macros/latex/logreq.tds.zip

*  Modified \defbibfilter syntax

   I've modified \defbibfilter such that it uses \ifboolexpr (from
   etoolbox.sty) instead of \ifthenelse (from ifthen.sty)
   internally. \ifboolexpr uses a slightly different syntax.
   For example, this definition:

   \defbibfilter{example}{%
     \( \type{book} \or \type{inbook} \)
     \and \keyword{abc}
     \and \not \keyword{x y z}
   }

   becomes:
   
   \defbibfilter{example}{%
     ( type=book or type=inbook )
     and keyword=abc
     and not keyword={x y z}
   }

   There is no need to update existing files since the old
   \ifthenelse syntax is still supported. \defbibfilter will
   implicitly 'translate' at definition time, if required.

*  Renamed option 'defernums' to 'defernumbers'

   The package option 'defernums' has been renamed to
   'defernumbers'.

*  Prefixed numerical citations, improved 'numeric' styles

   I've added three options to \printbibliography which are related
   to numerical citation schemes. They are called 'prefixnumbers',
   'resetnumbers', and 'omitnumbers'. Note that these options
   require that 'defernumbers' be set globally since they assign the
   numeric labels as the bibliography is generated (as opposed to at
   the beginning of the document body, as the data is read in).

   The new options enable the numeric styles to support prefixed
   numerical citations. The following may be accomplished by setting
   'prefixnumbers' for each prefixed subbibliography:

   REFERENCES
   Printed References
   [A1] ...
   [A2] ...
   Online References
   [B1] ...
   [B2] ...

   It is also possible to mix one non-prefixed subbibliography with
   one or more prefixed ones (set 'prefixnumbers' for each prefixed
   subbib and 'resetnumbers' for the non-prefixed one) or even to
   mix numerical citations with author-year or author-title
   citations (set 'omitnumbers' for all non-numerical subbibs).

   See the files '16-numeric-prefixed-1', '17-numeric-prefixed-2',
   and '18-numeric-hybrid' in the 'examples' subdirectory for
   practical examples.
   
   Adding support for prefixed numerical citations to existing
   numeric styles is trivial. The prefixes set in the argument to
   \printbibliography are available in the 'prefixnumber' field. If
   no prefix has been set, the field is undefined. Hence, in the cbx
   file, simply replace:

   \printfield{labelnumber}
   
   with:
   
   \printfield{prefixnumber}%
   \printfield{labelnumber}
   
   In the bbx file, the prefix needs to be incorporated in the
   definition of the bibliography environment. Replace:
   
   \defbibenvironment{bibliography}
     {\list
	{\printfield[labelnumberwidth]{labelnumber}}
	{...}
     {\endlist}
     {\item}

   with:
   
   \defbibenvironment{bibliography}
     {\list
	{\printtext[labelnumberwidth]{%
	   \printfield{prefixnumber}%
	   \printfield{labelnumber}}}
	{...}
     {\endlist}
     {\item}

   The numeric styles which ship with biblatex support prefixed
   labels out of the box.

*  New author-title style 

   There's a new style which rounds off the selection of
   author-title styles which ship with biblatex. It's called
   'authortitle-ticomp' and, as you can tell by the name, it's a
   variant of the 'authortitle-tcomp' style with an 'ibidem'
   feature.

*  Improved verbose styles, added 'citepages' option

   By popular request, all verbose styles now offer a 'citepages'
   option which deals with the 'pages' field in citations, in
   particular in cases in which the postnote is numerical.

   Note that style specific options are only mentioned briefly in
   the manual. They are discussed in greater details in the example
   file for the style (in the 'examples' subdirectory).

*  Long + short bibliography strings available

   Previous biblatex releases used to pick the long or the short
   version of the strings in the lbx files at file load time,
   depending on the setting of the global 'abbreviate' option.
   Starting with this release, both versions are kept in memory.

   In addition to \bibstring and friends, which print the long or
   the short version of the string (depending on the 'abbreviate'
   option), there are now commands like \biblstring and \bibsstring,
   which force the long or the short version, respectively.

   Note that the auxiliary macro \lbx@fromlang, which is used in lbx
   files, has been split up into \lbx@lfromlang and \lbx@sfromlang.

*  Automatically omit redundant language specifications

   Some style guides require that authors indicate the language of all
   foreign-language items in the bibliography. Biblatex's 'language'
   field (which is in fact a list) helps to deal with this
   requirement. If you are writing in more than one language and
   reuse the same set of bib files in different contexts, it is
   convenient to include this information whenever you add a new
   entry to a bib file. However, this leads to redundant
   information. It is rather odd to indicate "English" in the
   reference to an English book cited in an English article.

   The 'clearlang' option and \DeclareRedundantLanguages avoid such
   redundant language specifications. \DeclareRedundantLanguages
   maps languages as indicated in the 'language' field to the
   language identifiers of the babel package. Given this
   information, biblatex can omit the language if it matches the
   language of the document. Technically, this means that the
   'language' field will appear as undefined to styles, so it is
   safe to simply use '\printlist{language}' in styles and let
   biblatex sort out the rest.

   The 'clearlang' option controls this feature globally.
   \DeclareRedundantLanguages mappings are typically included in lbx
   files (but may also be given on a per-document basis in the
   preamble). All standard lbx files which ship with biblatex now
   include such mappings and 'clearlang' is enabled by default.

RELEASE NOTES FOR VERSION 0.9a

This release adds a few changes and improvements which were
originally scheduled for biblatex 0.9 but somehow got lost in the
release process or didn't make it in time. Pay attention to all
changes tagged as "CAVEAT".

*  [CAVEAT] Prenote placement with \cite, \textcite [numeric/alphabetic]

   The placement of the prenote in all numeric and alphabetic styles
   has been modified.

   Old format:

   \cite[see][15]{key}		->	see [1, p. 15]
   \parencite[see][15]{key}	->	[see 1, p. 15]
   \textcite[see][15]{key}	->	see Doe [1, p. 15]
   
   New format:

   \cite[see][15]{key}		->	[see 1, p. 15]
   \parencite[see][15]{key}	->	[see 1, p. 15]
   \textcite[see][15]{key}	->	Doe [see 1, p. 15]
   
   As seen in the above examples, the new format places the prenote
   inside the brackets. If you want the prenote outside of the
   brackets, you can simply use:

   see \cite[15]{key}		->	see [1, p. 15]
   see \textcite[15]{key}	->	see Doe [1, p. 15]
   
   which is similar to the old format.
   
*  [CAVEAT] Prenote placement with \textcite [authoryear/authortitle]

   The placement of the prenote in all authoryear and authortitle
   styles has also been modified.

   Old format:

   \cite[see][15]{key}		->	see Doe 2010, p. 15
   \parencite[see][15]{key}	->	(see Doe 2010, p. 15)
   \textcite[see][15]{key}	->	see Doe (2010, p. 15)
   
   New format:

   \cite[see][15]{key}		->	see Doe 2010, p. 15
   \parencite[see][15]{key}	->	(see Doe 2010, p. 15)
   \textcite[see][15]{key}	->	Doe (see 2010, p. 15)
   
   The new format places the prenote inside the brackets. If you
   want the prenote outside of the brackets, you can simply use:

   see \textcite[15]{key}	->	see Doe (2010, p. 15)
   
   which is similar to the old format.

*  [CAVEAT] Introducing \defbibenvironment

   I'm phasing out the thebibliography/theshorthands environments
   and the corresponding commands \thebibitem and \thelositem.
   Essentially, the old way of controlling the (high-level) layout
   of the bibliography and the list of shorthands is replaced by a
   new mechanism based on \defbibenvironment. It's not radically
   different. For example, where a style has code like this:

   \renewenvironment*{thebibliography}
     {\list{}{...}}
     {\endlist}
   \renewcommand*{\thebibitem}{\item}
   
   \renewenvironment*{theshorthands}
     {\list{}{...}}
     {\endlist}
   \renewcommand*{\thelositem}{\item}
   
   replace it with:
   
   \defbibenvironment{bibliography}
     {\list{}{...}}
     {\endlist}
     {\item}
   
   \defbibenvironment{shorthands}
     {\list{}{...}}
     {\endlist}
     {\item}
   
   The new system is similar to \defbibheading: you use
   \defbibenvironment to define the environment and the new 'env'
   option of \printbibliography and \printshorthands to select it.
   The point of the new system is that you can have different styles
   for different (partial) bibliographies in a document. It's also
   conceptually in line with \defbibheading now.

   I've added some compatibility code to biblatex. Older styles
   using thebibliography/thebibitem and theshorthands/thelositem
   will continue to work until you use \defbibenvironment for the
   first time to modify the default definition. From that point on,
   the old environment will be ignored.

   Updating old styles is of course highly recommended. The
   compatibility code may be removed from biblatex after 1.0. The
   standard styles have been updated to use the new syntax. They
   will ignore any changes based on the old definitions.

*  Added \printbibheading

   \printbibheading is the heading part of \printbibliography. This
   is useful for subdivided bibliographies. E.g., instead of

   \chapter{\bibname}
   \printbibliography[heading=subbibliography,type={...},...]
   \printbibliography[heading=subbibliography,type={...},...]
   ...

   you use:

   \printbibheading % = \printbibheading[heading=bibliography]
   \printbibliography[heading=subbibliography,type={...},...]
   \printbibliography[heading=subbibliography,type={...},...]
   ...
   
   The point is that the overall heading is in sync with the
   subheadings.

*  Smart parentheses and brackets

   This release adds a new 'parenthesis tracker', i.e., if
   parentheses and/or brackets are nested, biblatex alternates
   between parentheses and brackets, depending on the nesting level.
   E.g.:

   \mkbibparens{text \mkbibparens{text} text}
   \mkbibbrackets{text \mkbibbrackets{text} text}
   
   yields
   
   (text [text] text)
   [text (text) text]

   This also works with \bibopenparen + \bibcloseparen and
   \bibopenbracket + \bibclosebracket as well as the new citation
   commands \parentext and \brackettexts.

*  Open ended date ranges

   There's one thing I forgot to mention when introducing the new
   date fields in the previous release. It is possible to specify an
   open ended range by leaving the end date blank:
   
   1988/1992		   = 1988-1992
   1988/		   = 1988-today
   1988			   = 1988
   
   Open ended ranges are represented by a defined but empty
   "endyear" component on the level of the style interface:

   date = {1988}	->	day		= undefined
   				month		= undefined
				year		= "1988"
   				endday		= undefined
   				endmonth	= undefined
   				endyear		= undefined
   
   date = {1988/}	->	day		= undefined
   				month		= undefined
				year		= "1988"
   				endday		= undefined
   				endmonth	= undefined
   				endyear		= EMPTY
   
   date = {1988/1992}	->	day		= undefined
   				month		= undefined
				year		= "1988"
   				endday		= undefined
   				endmonth	= undefined
   				endyear		= "1992"
   
   All high-level date commands like \printdate detect and handle
   open ended ranges autoamtically. Use

   \iffieldequalstr{endyear}{}
     {true}
     {false}
   
   if you want to test for open ended ranges explicitly.

*  Extended language support: Finnish

   This release comes with a finnish.lbx file. The translations
   have been contributed by Hannu V\"ais\"anen.

*  Updated language support: Greek

   This release comes with an updated greek.lbx file contributed
   by biblatex user Prokopis.

*  Updated examples, some updates in the manual

   Some of the examples which ship with biblatex were not completely
   up to date in the 0.9 release. They have been updated now. I've
   also corrected/added a few minor points in the manual.

RELEASE NOTES FOR VERSION 0.9

Several features on my 'to do' list have been deferred because they
introduce backwards compatibility issues. I wanted to implement them
all in one shot so that style authors only need to update their
styles once. This release finally takes that step. In other words:

	THIS VERSION IS NOT FULLY BACKWARDS COMPATIBLE!

Style authors will most likely need to update their styles to make
them work with this release. You will find detailed hints about what
needs to be updated in the release notes below. Changes which raise
backwards compatibility issues are tagged as "CAVEAT". Users who
rely on styles found in:

	macros/latex/exptl/biblatex-contrib/

on CTAN are advised to stick with biblatex 0.8 for the time being
and defer the upgrade to 0.9 until the styles they use have been
updated. If you have accidentally installed this release and find
out that you need to downgrade, you can get the previous biblatex
release from the project page on Sourceforge:

	http://sourceforge.net/projects/biblatex/files/

CTAN always carries the latest version.

I have received many inquiries about why biblatex is not yet
included in TeX Live and MikTeX and why I was maintaining that nasty
do-not-distribute clause. There seems to be some confusion about
what the term 'stable' means. The point typically brought up was
that biblatex runs well and is thus 'stable'. Well, in that sense,
biblatex has been pretty stable since version 0.1. However, it has
not been stable in the sense of maintaining backwards compatibility
since version 0.1. This release, however, is finally considered
suitable for wider distribution.

*  Distributing biblatex

   This release should be a reasonable target for TeX distributions,
   even though biblatex is still tagged as 'beta' and located in the
   experimental branch on CTAN. I will make the move to the stable
   branch when it hits 1.0. Given the number of changes and new
   features in this release, it's quite possible that some bugs
   slipped in. If you start packaging biblatex now, be prepared for
   some maintenance releases (i.e., 0.9a, 0.9b, etc.) in the
   upcoming weeks.

   A few more notes to distributors:
   
   - See the README file for requirements and dependencies.
   
   - The *.csf files in doc/resources go into TEXMF/bibtex/csf/.
     The names will not conflict with the *.csf files shipping with
     biblatex8. Biblatex uses names like latin1.csf while biblatex8
     uses 88591lat.csf.

   - It would be nice to have two copies of biblatex-examples.bib
     installed, one in TEXMF/doc for users looking for documentation
     in the usual place and a second copy in TEXMF/bibtex/bib/.
     Since users are encouraged to use this database in minimal
     examples, it's handy to have it readily available.

   - Biber is the next-generation backend of biblatex. It would be
     very welcome to have a Biber package as well. See the project
     page on http://biblatex-biber.sourceforge.net/ for details.
     Note that biblatex currently supports three backends: BibTeX
     (i.e., traditional 7-bit BibTeX), bibtex8, and Biber.
     BibTeX/bibtex8 support will be discontinued in biblatex 2.x.

   - A package for the contributed styles found in

	macros/latex/exptl/biblatex-contrib/

     is also welcome because some of these styles are already very
     popular. However, wait for the styles in to be updated to
     biblatex 0.9 before packaging them.

*  Contributing to biblatex

   Localizing biblatex heavily depends on user contributions. I've
   added a new tracker named 'Contrib' to the project page which
   will hopefully help to organize this process:

   http://sourceforge.net/tracker2/?group_id=244752
   
   There is one item for each language supported by biblatex. The
   status of the item reflects the localization status of the
   corresponding language. If the status is "open", the localization
   module is incomplete or in need of peer review by a native
   speaker. If the status is "closed", the module is complete and
   there are no known issues (you can still post comments to closed
   items and upload files if you want to suggest improvements).

   If you have updates for a language module, upload the updated
   *.lbx file in the appropriate thread. If you want to contribute
   support for a new language, simply open a new tracker item.

*  Updating existing biblatex styles

   As usual, the full list of changes is found in the biblatex
   manual. The release notes you are just reading mention all
   changes which are not backwards compatible. Pay attention to all
   "CAVEAT" points.

   When you update your style, it may be a good idea to add some
   code which checks that the style is running under biblatex 0.9 or
   later. You can use the standard LaTeX \@ifpackagelater test for
   this purpose. Here's some sample code (where 'YYYY/MM/DD' needs
   to be replaced with the real release date of biblatex 0.9 and
   'MYSTYLE' with the name of your style):

   \@ifpackagelater{biblatex}{YYYY/MM/DD}
     {}
     {\PackageError{biblatex}
	{Outdated 'biblatex' package}
	{The 'MYSTYLE' style requires biblatex v0.9 or later.\MessageBreak
	 You are using: '\csuse{ver@biblatex.sty}'.\MessageBreak
	 This is a fatal error. I'm aborting now.}%
      \endinput}

   Since \@ifpackagelater is in fact a 'greater than or equal' test,
   you can simply replace YYYY/MM/DD with the date found in the
   \ProvidesPackage declaration at the top of biblatex.sty.

*  New Biber interface

   Biber is the next-generation database backend of biblatex. It is
   available from:

	http://biblatex-biber.sourceforge.net/
   
   If you decide to use Biber, set 'backend=biber' (preferably in
   biblatex.cfg) and upgrade to Biber 0.5 or higher. Biblatex 0.9
   uses a new interface to talk to Biber hence you must specify
   backend=biber and use the latest Biber version.
   
   Note that biblatex and Biber use separate bug trackers. If you're
   using Biber and find a bug, try your example with a different
   backend. If the problem persists, report it as a biblatex bug. If
   not, report it as a Biber bug.

*  [CAVEAT] Revised and expanded date fields

   All date fields have been expanded such that they support ranges
   in 'start/end' format. Of course it is still possible to specify
   a single date and you may also truncate dates at the end. In sum,
   the following examples are all valid date specifications:

   1850
   1967-02
   2009-01-31
   1988/1992		   <- this is a range (from 1988 until 1992)
   2002-01/2002-02
   1995-01-31/1995-02-05

   This format is supported by the following fields: 'date',
   'origdate', 'eventdate', 'urldate' (for the sake of consistency
   in the case of 'urldate'). I have also removed or modified some
   of the fields for date components (e.g., 'day', 'urlday', etc.):

   * [CAVEAT] The field 'day' has been removed. Use the 'date' field
     instead to specify full dates.

   * [CAVEAT] The field 'year' is no longer a range field but a
     literal field. Year ranges should be given in the 'date' field
     instead.

   * [CAVEAT] I've removed the field 'origyear' and added 'origdate'
     as a replacement. 'origdate' is similar to 'date', i.e., it
     takes a date specification in 'yyyy-mm-dd' format and supports
     date ranges in 'start/end' format.

   * [CAVEAT] The fields 'urlday', 'urlmonth', 'urlyear' have been
     removed. Use 'urldate' instead.

   * I've added an 'eventdate' field. This is also a date range
     field.

   USERS: You will probably need to update your bib files. The
   changes which are most likely to cause incompatibilities are the
   removal of the 'day' field and the modification of 'year'. Use
   'date' instead were required. The removal/renaming of 'origyear'
   may also require a few changes. I doubt than many users are
   specifying access dates using urlday+urlmonth+urlyear. Since
   'urldate' has been extended but not modified in an incompatible
   way, this change should be fairly transparent.

   AUTHORS: The way the date components are made available to styles
   has not changed. For example, the components of the publication
   date are available as day/month/year no matter whether they were
   given in the 'date' field or in 'year' and 'month'. In sum,
   'date'-like fields are split up as follows:

   *.bib file		style interface
   ------------------   ------------------
   
   date			day
   			month
			year
   			endday
   			endmonth
   			endyear

   origdate		origday
   			origmonth
			origyear
   			origendday
   			origendmonth
   			origendyear

   eventdate		eventday
   			eventmonth
			eventyear
   			eventendday
   			eventendmonth
   			eventendyear

   urldate		urlday
   			urlmonth
			urlyear
   			urlendday
   			urlendmonth
   			urlendyear

   Fields related to date components which are not available are
   undefined. E.g.:
   
   date = {1988}	->	day		= undefined
   				month		= undefined
				year		= "1988"
   				endday		= undefined
   				endmonth	= undefined
   				endyear		= undefined
   
   urldate = {2009-01-31}
			->	urlday		= "31"
   				urlmonth	= "01"
				urlyear		= "2009"
   				urlendday	= undefined
   				urlendmonth	= undefined
   				urlendyear	= undefined
   
   origdate = {2002-01/2002-02}
   			->	origday		= undefined
   				origmonth	= "01"
				origyear	= "2002"
   				origendday	= undefined
   				origendmonth	= "02"
   				origendyear	= "2002"
   
   eventdate = {1995-01-31/1995-02-05}
   			->	eventday	= "31"
   				eventmonth	= "01"
				eventyear	= "1995"
   				eventendday	= "05"
   				eventendmonth	= "02"
   				eventendyear	= "1995"

   If there is no 'date' field, biblatex will consider the legacy
   fields 'month' and 'year'. E.g.:

   year = {1988}	->	day		= undefined
   				month		= undefined
				year		= "1988"
   				endday		= undefined
   				endmonth	= undefined
   				endyear		= undefined
   
   year  = {1993}	->	day		= undefined
   month = {5}			month		= "05"
				year		= "1993"
   				endday		= undefined
   				endmonth	= undefined
   				endyear		= undefined
   
   When printing a single date component in a style, use the field
   names listed in the 'style interface' column above (e.g.,
   "\printfield{origmonth}"). There is no 'date' field on the style
   level. However, there is normally no need to fiddle with the date
   component fields directly. See the next point for details.
   
*  [CAVEAT] Modified date interface

   The high-level date interface has been revised. The commands
   \bibdate and \biburldate have been renamed and improved. Here's a
   list of the new high-level commands:

      \printdate
      \printurldate
      \printorigdate
      \printeventdate
   
   These commands print the respective date, handle ranges, and
   localize the output format. For example, if the publication date
   is a plain year, then \printdate will only print the year. If
   it's a full date, it will print all date components which are
   available. If it's a range, it will format the range. The
   high-level commands need not be wrapped in a \printtext command.

   USERS: The change is relevant for style authors only. The standard
   styles have been updated.

   AUTHORS: Where your style has code like:
   
      \bibdate
      or \printtext{\bibdate}

   and

      \biburldate
      or \printtext{\biburldate}

   use
   
      \printdate

   and

      \printurldate

   instead. Use \printorigdate and \printeventdate to handle the new
   'origdate' and 'eventdate' fields.

*  [CAVEAT] Modified 'labelyear'

   Adding support for date ranges has prompted a refactoring of the
   'labelyear' mechanism. Essentially, I've renamed the existing
   'labelyear' field to 'extrayear' and re-added a new 'labelyear'
   field with a different role. The 'maxlabelyear' counter has been
   renamed to 'maxextrayear'. The new mechanism is consistent with
   labelalpha/extraalpha, e.g., 'labelyear' holds the bare year (or
   year range) and 'extrayear' indicates the extra letter used for
   disambiguation (as an integer).

   The new 'labelyear' field holds the year components of the 'date'
   field, readily formatted using \bibdatedash as a range separator.
   Identical year components in a date range are detected
   automatically. If there is no 'date' field, the 'year' field will
   be considered instead. Here are some examples:

   date = {2003}
		->	labelyear = "2003"

   date = {1995-01-31/1995-02-05}
		->	labelyear = "1995"

   date = {1995/1998}
		->	labelyear = "1995\bibdatedash 1998"

   date = {1998-12-27/1999-01-03}
		->	labelyear = "1998\bibdatedash 1999"

   year = {2005}
		->	labelyear = "2005"

   If there is neither a 'date' nor a 'year' field then 'labelyear'
   will be undefined. 

   USERS: The change is relevant for style authors only. The standard
   styles have been updated.

   AUTHORS: Where your style has code like:

      \printfield{year}%
      \printfield{labelyear}%

   use this:

      \printfield{labelyear}%
      \printfield{extrayear}%

   instead. If you want to concatenate 'extrayear' and the (full)
   publication date in the bibliography, use:

      \printdateextra

   \printdateextra is similar to \printdate but incorporates the
   'extrayear' in the date specification.

*  [CAVEAT] Revised date localization

   The way date specifications are localized in lbx files has been
   refactored. Only \DeclareBibliographyExtras declarations in *.lbx
   files are affected by this change. 

   Basically, the commands:
   
	\bibdatelong
	\bibdateshort
	\biburldatelong
	\biburldateshort

   have been removed and are replaced by the generic commands:

	\mkbibdatelong
	\mkbibdateshort
   
   The new commands take three arguments (three field names) and
   arrange them as required by the respective language/locale. The
   new system is more easily extensible.

   USERS: The change is relevant for style authors only. The standard
   styles have been updated.

   AUTHORS: Even though this point is tagged as 'caveat', it's
   rather unlikely that it will cause any trouble. The change is
   only relevant for style authors who a) ship custom lbx files (or
   users who have adapted some of the stock lbx files) and b) have
   modified the 'extras' (as opposed to inheriting them form the
   stock modules).

*  Added/extended package options for date formats

   There are now four date formats to choose from: short, long,
   terse, comp. terse and comp are like short and long but render
   date ranges in a more compact format.

*  Support for multiple editorial roles

   This release adds three additional fields (data type: name list)
   for secondary editors: 'editora', 'editorb', 'editorc'. There are
   also matching 'editor*type' fields. Biblatex now supports up to
   four distinct editoral roles, e.g.:

   editor      = {Eddy Editor}
   editora     = {Freddy Founder}
   editoratype = {founder}
   editorb     = {Cory Continuator}
   editorbtype = {continuator}
   editorc     = {Rudy Redactor}
   editorctype = {redactor}
   
   USERS: The new fields are fully supported by the standard styles.

   AUTHORS: The catch-all 'byeditor+others' macro in biblatex.def
   has been updated to incorporate the new fields. The 'byeditor'
   macro will consider them as well. There's also a dedicated
   'byeditorx' macro which processes the 'editor[a-c]' lists. In
   other words: if your style relies of the ready-made macros in
   biblatex.def, it will consider the new fields automatically. If
   not, have a look at biblatex.def to get an idea of how to
   incorporate them in your style.

*  [CAVEAT] 'redactor' now an editorial role

   The field 'redactor' has been removed because the redactor is now
   an editorial role, i.e.:
   
   editor      = {Eddy Editor}
   redactor    = {Rudy Redactor}
   
   becomes:
   
   editor      = {Eddy Editor}
   editora     = {Rudy Redactor}
   editoratype = {redactor}
   
   USERS: You may need to update some of your bib files.
   
   AUTHORS: The 'byredactor' macro has been removed from
   biblatex.def because this role is now handled by the
   editor-related macros mentioned above.

*  [CAVEAT] Revised lbx files, simplyfied role concatenation

   I've simplyfied the role concatenation somewhat. Concatenation is
   now only supported if the primary role is 'editor' (i.e., if the
   'editortype' field is undefined or holds the string 'editor'). It
   turned out to be too painful to support concatenation for all
   kinds of editorial roles (compiler, redactor, founder, etc.) in
   all languages. This will also make the lbx files much more
   readable and the whole apparatus easier to understand.

   It's all a bit difficult to describe but fairly easy to grasp if
   you take a look at the strings in the lbx files. Essentially,
   biblatex 0.8 would define fragmentary type* and bytype* snippets:

   typeeditor       = {{editor}{...}},
   typecompiler     = {{compiler}{...}},
   bytypeeditor     = {{edited}{...}},
   bytypecompiler   = {{compiled}{...}},

   The full strings would then be assembled by auxiliary macros
   which insert the type* and bytype* snippets into the string
   definitions:

   editor           = {{\lbx@typeeditor}{...}},
   editortr         = {{\lbx@typeeditor\ and translator}{...}},
   byeditor         = {{\lbx@bytypeeditor\ by}{...}},
   byeditortr       = {{\lbx@bytypeeditor\ and translated \lbx@fromlang\ by}{...}},

   These are the simplified strings in biblatex >=0.9:

   editor           = {{editor}{...}},
   compiler         = {{compiler}{...}},
   editortr         = {{editor and translator}{...}},
   byeditor         = {{edited by}{...}},
   bycompiler       = {{compiled by}{...}},
   byeditortr       = {{edited and translated \lbx@fromlang\ by}{...}},

   The old scheme was admittedly more elegant but it simply didn't
   work quite right in all languages.

*  Notes on the basic structure of role processing in biblatex.def

   The following hints may help you to find your way around the
   ready-made macros and definitions in biblatex.def.
   
   Essentially, a role like 'editor' may be expressed as a function
   or as an action, e.g.:

   Bernard Bookmaker, editor, Title, ...	[function]
   ... Title, edited by Bernard Bookmaker	[action]
   
   Roles which are related to supplementary material in a book, such
   as a commentary or annotations, may also be expressed as an
   object:

   ... Title, annotated by Edward Expert	[action]
   ... Title, with annotations by Edward Expert	[object]
   
   In biblatex.def, the names of the ready-made macros correspond to
   name formats and localization keys:

   Macros like editor/translator print a role as a function.
   They use
   - name formats like editor/translator and
   - strings like editor/translator.

   Macros like byeditor/bytranslator print a role as an action.
   They use
   - name formats like byeditor/bytranslator and
   - strings like byeditor/bytranslator.

   Macros like withcommentator/withannotator print a role as an
   object. They use
   - name formats like withcommentator/withannotator and
   - strings like withcommentator/withannotator.

   For example, the bibmacro 'editor' essentially boils down to the
   following code:

   \printnames{editor}    = \printnames[editor]{editor}
   \bibstring{editor}
   
   The 'byeditor' macro boils down to:
   
   \bibstring{byeditor}
   \printnames[byeditor]{editor}
   
   and the 'withannotator' macro to:
   
   \bibstring{withannotator}
   \printnames[withannotator]{annotator}
   
   The real code is more complex because it needs to take
   punctuation and additional parameters into account (such as the
   'editortype' field, role concatenation, etc.) but this is the
   basic idea.

*  [CAVEAT] Revised indexing controls

   The way the indexing facilities are configured has been revised
   and modified. In previous releases, the 'indexing' option would
   enable and disable commands like \indexfield on a low level.
   Starting with this release, the indexing commands are always
   operational. The 'indexing' option controls two new tests called
   \ifciteindex and \ifbibindex which should be used as explicit
   tests.

   In the standard styles, indexing is handled by two bibmacros
   called 'citeindex' and 'bibindex'. The old definition was:

   \newbibmacro*{citeindex}{%
     \indexnames{labelname}%
     \indexfield{indextitle}}

   \newbibmacro*{bibindex}{%
     \indexnames{labelname}%
     \indexfield{indextitle}}

   The new one is:

   \newbibmacro*{bibindex}{%
     \ifbibindex
       {\indexnames{labelname}%
	\indexfield{indextitle}}
       {}}

   \newbibmacro*{citeindex}{%
     \ifciteindex
       {\indexnames{labelname}%
	\indexfield{indextitle}}
       {}}

   The point of this change is to make commands like \indexfield
   useable in a more flexible way, regardless of the setting of the
   'indexing' option. Since they are always operational now, you can
   use them for tasks which require access to the bibliographic data
   but do not print anything, such as setting page headers or
   writing data to some external file.

   It is not advisable to use commands like \printfield for this
   purpose because they interface with the punctuation tracker. Use
   \printfield if you really want to print a field and \indexfield
   for other tasks, indexing obviously being the most common
   application.

   USERS: From a user's point of view, the new mechanism works like
   the old one.

   AUTHORS: Standard citation commands like \cite should always use
   the above bibmacros (citeindex/bibindex) anyway so no change is
   required. You may need to incorporate the \ifciteindex and
   \ifbibindex tests into text citation commands (i.e., commands
   like \citeauthor and \citetitle) if you have redefined the
   standard ones or provide additional ones. All predefined commands
   in biblatex.def have been updated.

*  Fallback drivers

   Biblatex now supports the fallback entry type "*". If you use the
   asterisk as the type argument of \DeclareBibliographyDriver or
   \DeclareBibliographyAlias, the driver or alias will be used as a
   fallback if no specific driver for a certain type has been
   defined. E.g.:

   \DeclareBibliographyDriver{*}{<driver code>}
   \DeclareBibliographyAlias{*}{misc}
   
*  Manual language selection

   By default, biblatex detects the babel package if it is loaded
   and can adjust to the selected babel language automatically.
   Using the new 'language' package option, you can now force
   biblatex to load support for a certain language. This will
   implicitly disable babel support and on-the-fly language
   switching. The 'babel' package option is ignored in this case.

   This is mainly useful if you can't use babel for some reason.
   Using the automatic babel interface is usually preferable.

*  Improved \defbibheading, introducing 'title' option

   I've extended \defbibheading and added a 'title' option to
   \printbibliography and \printshorthands. These extensions are
   mainly useful for subdivided bibliographies.
   
   This is best explained by example. With previous biblatex
   releases, subdivided bibliographies required one heading
   definition for each subsection since the heading definition
   included both the format and the text of the heading:

   \defbibheading{books}{\section*{Books}}
   \defbibheading{articles}{\section*{Articles}}
   \defbibheading{online}{\section*{Online Resources}}

   \chapter{\bibname}
   \printbibliography[type=book,heading=books]
   \printbibliography[type=article,heading=articles]
   \printbibliography[type=online,heading=online]

   When using the extended syntax of \defbibheading, the heading
   definition only controls the format but the title may be
   specified by using the 'title' option of \printbibliography and
   \printshorthands. The value of this option will be passed to the
   heading definition as paramater #1. The new optional argument of
   \defbibheading defines the default text which is passed as #1 if
   there is no 'title' option.

   \defbibheading{subdiv}[\bibname]{\section*{#1}}

   \chapter{\bibname}
   \printbibliography[type=book,heading=subdiv,title={Books}]
   \printbibliography[type=article,heading=subdiv,title={Articles}]
   \printbibliography[type=online,heading=subdiv,title={Online Resources}]

   Since biblatex provides a predefined heading called
   'subbibliography' you don't even need \defbibheading in this
   case. You can simply use 'heading=subbibliography', specify
   'title={...}', and rely on the predefined heading definitions
   provided by biblatex which support the standard LaTeX classes,
   the KOMA-Script classes, and the memoir class automatically.

   The extensions are backwards-compatible. It is still possible to
   use the parameterless form of \defbibheading. Such headings will
   work as expected but the 'title' option of \printbibliography and
   \printshorthands will obviously be ignored in this case.

*  More options for \printshorthands, subdivided LOS

   I've added more options to \printshorthands. It now supports the
   same options as \printbibliography. This is useful for a
   subdivided list of shorthands.

*  Improved author-title and author-year citation styles

   I've improved the compact citation styles authortitle-comp,
   authortitle-icomp, authortitle-tcomp, and authoryear-comp. The
   compact format is now supported in multicite commands as well.

*  New verbose-trad3 style

   This style is similar in concept to verbose-trad2 with a slightly
   different treatment of ibidem and op. cit.

*  Eprint support for Google Books/JSTOR/PubMed

   I've added eprint handlers for Google Books, JSTOR, and PubMed.
   This means that, instead of

   url	= {http://books.google.com/books?id=XXu4AkRVBBoC},

   you may also use:

   eprint     = {XXu4AkRVBBoC},
   eprinttype = {googlebooks},

   PubMed support is similar. Instead of specifying the full URL
   (where <pmid> is the unique and stable PubMed ID):

   url	= {http://www.ncbi.nlm.nih.gov/pubmed/<pmid>},

   you may also say:

   eprint     = {<pmid>},
   eprinttype = {pubmed},

   By default, this will be printed as "PMID: <pmid>" in the
   bibliography. If hyperref support is enabled, the <pmid> will be
   a clickable link to PubMed.

   JSTOR support works exactly like PubMed. When using JSTOR's
   export feature to export citations in BibTeX format, JSTOR uses
   the 'url' field by default (where <number> is a unique
   and stable identifier):

   url = {http://www.jstor.org/stable/<number>}

   While this will work as expected, full URLs tend to clutter the
   bibliography. You may also use:

   eprint     = {<number>},
   eprinttype = {jstor},
   
   to get the more readable "JSTOR: <number>" format which also
   supports hyperlinks (the <number> becomes a clickable link).

*  More styles for compressed back references

   See 'backrefstyle' in the changelog.

*  Improved endnotes, added \pagenote support

   The \pagenote command, which is provided by the pagenote package
   and the memoir class, is now supported by \mkbibendnote.
   \mkbibendnote prioritizes as follows:

   1) check for \endnote and use it if available
   2) check for \pagenote and use it if available
   3) issue a warning and fall back to \footnote
   
   There is also a new 'notetype' option which you may use to
   convert footnotes to endnotes and vice versa (provided that
   they are generated with \mkbibfootnote and \mkbibendnote,
   respectively; the standard \footnote and \endnote commands will
   not be modified).

*  Compatibility with UCS package

   Previous releases of biblatex were comptabile with inputenc's
   standard UTF-8 module but wouldn't work with the ucs package. This
   release also supports ucs. XeLaTeX's native UTF-8 support should
   also work fine. If you're using UTF-8 encoding you should be
   using Biber instead of BibTeX as a backend.

*  Dutch support

   This release comes with a dutch.lbx file. The translations have
   been contributed by Alexander van Loon.

*  Greek support

   As of this release, biblatex speaks Greek, too. The translations
   have been contributed by Apostolos Syropoulos. Note that the
   greek.lbx file requires UTF-8 support. Biblatex generally works
   with

   1) LaTeX and inputenc:

      \usepackage[utf8]{inputenc}

   2) LaTeX and inputenc+ucs:

      \usepackage{ucs}
      \usepackage[utf8x]{inputenc}

   3) XeLaTeX

   Since inputenc's standard utf8 module has no glyph mappings for
   Greek, this leaves Greek users with a choice of 2) or 3). If you
   choose 2), make sure to preload the Greek Unicode range:

   \usepackage{ucs}
   \usepackage[utf8x]{inputenc}
   \PreloadUnicodePage{3}
   
   You may also use \PrerenderUnicode if you prefer that. See the
   section about "Known problems" in the UCS (!) manual for
   explanation. You may also need to load additional packages which
   set up Greek fonts. As a rule of thumb, any setup which works with
   regular Greek documents should also work with biblatex.

   However, there is one fundamental limitation. As of this writing,
   biblatex has no support for mixing scripts. Bibliographies in
   Greek should work fine, provided that you use Biber as a backend,
   but English and other titles in the bibliography may be rendered
   in Greek letters.
   
   The problem with using LaTeX and macro-level UTF-8 support via
   inputenc/ucs is that processing still happens in an 8-bit
   environment. The inputenc/ucs packages can decode all of UTF-8,
   but in order to typeset Unicode all input needs to be mapped to
   segments of 256 glyphs each because LaTeX only supports 8-bit
   output encodings. If you need multi-script bibliographies,
   XeLaTeX is the only sensible choice since XeTeX uses Unicode all
   the way from input to output.

RELEASE NOTES FOR VERSION 0.8i

This is a bug fix release. There are no new features and the manual
has not been updated.

RELEASE NOTES FOR VERSION 0.8h

This is a bug fix release. There are no new features and the manual
has not been updated.

RELEASE NOTES FOR VERSION 0.8g

This is a bug fix release. There are no new features and the manual
has not been updated.

RELEASE NOTES FOR VERSION 0.8f

This is a bug fix release. There are no new features and the manual
has not been updated.

RELEASE NOTES FOR VERSION 0.8e

*  Improved concatenation of roles

   The standard styles (and the ready-made macros in biblatex.def)
   have been supporting the concatenation of (editorial and other)
   roles for some time. For example, if the editor and the
   translator of a book are the same person, the roles are
   concatenated and the name is given only once:

   Author, Title, ed. and trans. by Editor, ...
   
   However, if there is no author and the editor moves to the first
   position, the name was, until now, printed twice:

   Editor, ed., Title, trans. by Editor, ...
   
   The improved bibliography styles now support concatenation in
   this case as well:

   Editor, ed. and trans., Title, ...
   
   Since the macros which handle the concatenation are defined in
   biblatex.def, adding this feature to existing styles is simple.
   Where your style has code like:

   \usebibmacro{editor}
   \usebibmacro{author/editor}
   
   simply replace that with:
   
   \usebibmacro{editor+others}
   \usebibmacro{author/editor+others}
   
   The improvements also cover additional roles of the translator.
   To take advantage of that, replace code like:

   \usebibmacro{translator}
   \usebibmacro{author/translator}
   \usebibmacro{author/editor/translator}
   
   with:
   
   \usebibmacro{translator+others}
   \usebibmacro{author/translator+others}
   \usebibmacro{author/editor+others/translator+others}

   On a related note, it is safe to use code like this:

   \usebibmacro{editor+others}%
   \newunit
   \printfield{title}%
   \newunit
   \usebibmacro{byeditor+others}%
   
   because macros like 'editor+others' macro will use \clearname to
   clear all name lists already processed. For example, if the
   'editor+others' macro prints 'ed. and trans.', it locally clears
   the 'editor' and 'translator' lists afterwards. The
   'byeditor+others' macro will not print them a second time. It
   simply takes care of any names/roles not covered yet
   (commentator, annotator, etc.), if there are any left.

*  Added some new localization keys, removed some existing ones

   The above concatenation requires a whole set of new localization
   keys (see the manual for details). These keys are initialized by
   default. English and German translations are already available,
   the other localization modules need an update.
   
   The new strings are also useful for styles which use the 'first
   position' format in all cases, i.e., styles which print:

   Editor, ed. and trans., Title, ...
   Author, Title, in: Editor, ed. and trans., Book, ...
   
   instead of:

   Editor, ed. and trans., Title, ...
   Author, Title, in: Book, ed. and trans. by Editor, ...

   Since the number of localization keys keeps increasing, I've
   removed some of the rarely used 'country...', 'patent...', and
   'patreq...' keys. Only about half a dozen keys in each of these
   groups is left, mainly to illustrate the underlying scheme.

*  New auxiliary macros for ordinals

   In the past, some localization modules have been redefining the
   field formats 'edition' and 'series' because they require
   ordinals and need to be adapted to use either \mkbibmascord or
   \mkbibfemord if ordinals are gender specific in the respective
   language. However, having lbx files redefine field formats is
   rather intrusive.

   Starting with this release, there's a new, non-intrusive
   solution. Biblatex provides two new macros, \mkbibordedition and
   \mkbibordseries, which should be used in the respective field
   formats. Where required (typically in Romanic languages), these
   macros are redefined by localization modules such that they point
   to \mkbibmascord or \mkbibfemord.

*  Extended language support: Brazilian Portuguese, Swedish revised

   This release adds support for Brazilian Portuguese and
   preliminary support for Portuguese/Portugal. The translations
   have been contributed by Augusto Ritter Stoffel.

   The Portuguese support in portuguese.lbx is mostly inherited from
   brazilian.lbx. This file needs review by a native speaker from
   Portugal.

   I've also added some revised Swedish translations which Per
   Starb\"ack sent in some time ago to the Swedish module.

*  German localization now using "Hrsg."

   By popular request, the German module now uses the abbreviations
   "Hrsg./hrsg. von" instead of "Hg./hg. von". Both forms are valid,
   but most users seem to prefer the more traditional "Hrsg".

RELEASE NOTES FOR VERSION 0.8d

*  Biber beta release

   Biber is a BibTeX replacement written in Perl. It features
   Unicode support and has been designed with the requirements of
   biblatex in mind. See:

   http://biblatex-biber.sourceforge.net/
   
   for details. Beta testers are invited to try it out and report
   any bugs on Biber's SourceForge project page.

*  Package option 'bibtex8' superseded by 'backend'

   The package option 'bibtex8' is superseded by a new option called
   'backend'. The old option is not mentioned in the manual any
   more, but it is still supported for the sake of backwards
   compatibility. In new documents, however, replace it as follows:

   bibtex8=false -> backend=bibtex  (default setting, omissible)
   bibtex8=true	 -> backend=bibtex8
   
RELEASE NOTES FOR VERSION 0.8c

*  Added 'idem' tracker

   I've added an 'idem' tracker which is similar in concept to the
   'ibidem' tracker except that it checks for recurrent author/
   editor names. See the 'idemtracker' package option and the
   \ifciteidem test in the manual.

RELEASE NOTES FOR VERSION 0.8b

*  Added 'usetranslator' option

   The 'usetranslator' option is similar in concept to 'useauthor'
   and 'useeditor' hence usage should be fairly obvious. Style
   authors should note that styles derived from the standard ones
   will support this option if you replace

   \usebibmacro{author}
   \usebibmacro{author/editor}

   with
   
   \usebibmacro{author/translator}
   \usebibmacro{author/editor/translator}
   
   in all relevant drivers.

RELEASE NOTES FOR VERSION 0.8a

*  SourceForge bug/feature trackers

   The user base of biblatex has been growing steadily, reaching a
   point where managing bug reports and feature requests by a
   combination of private email messages and Usenet postings has
   become impractical. I've therefore set up a project page on
   SourceForge:

   http://sourceforge.net/projects/biblatex/

   There's no code on that site and most tools offered by
   SourceForge (like cvs/subversion, web-based forums, etc.) are
   currently disabled. The interesting thing are the trackers:

   http://sourceforge.net/tracker2/?group_id=244752
   
   I've set up two trackers, 'Bugs' for bug reports and 'Features'
   for feature requests. Please use these trackers to report bugs
   and submit feature requests.
   
   I've also added all open bug reports and feature requests sitting
   in my email inbox but I may have missed some messages posted in
   public forums. If you've reported anything which doesn't show up
   on the project page, you can now add it yourself by selecting the
   appropriate tracker and clicking on "Add new artifact".

*  Custom localization modules

   This release adds support for custom localization modules. The
   point is that styles may ship modified lbx files. See
   \DeclareLanguageMapping and the changelog for further hints.

*  Configurable punctuation tracker

   The behavior of punctuation commands like \addcomma is now
   configurable. See \DeclarePunctuationPairs in the manual.

*  Improved 'American-style' punctuation

   I've improved the 'American-style' punctuation feature.
   american.lbx now uses \DeclarePunctuationPairs to adapt the
   punctuation tracker and \mkbibquote supports nested quotes even
   if American punctuation is enabled.

*  Sentence case vs. title case

   By popular request, I've added a macro which converts a string to
   sentence case. See \MakeSentenceCase, \MakeSentenceCase*, and
   \DeclareCaseLangs in the manual. \MakeSentenceCase supports the
   BibTeX convention that anything wrapped in braces is not altered
   when changing the case.
   
   Style authors who make use of the bibmacros in biblatex.def should
   note the following. Saying:
   
   \DeclareFieldFormat{title}{\MakeSentenceCase{#1}}
   
   will not work as expected. That's because biblatex uses macros
   like this one by default:
   
   \newbibmacro*{title}{%
     \ifthenelse{\iffieldundef{title}\AND\iffieldundef{subtitle}}
       {}
       {\printtext[title]{%
	  \printfield[noformat]{title}%
	  \setunit{\subtitlepunct}%
	  \printfield[noformat]{subtitle}}%
	\newunit}%
     \printfield{titleaddon}}
   
   The 'title' format is applied by a \printtext command which
   encloses both the title and the subtitle. \MakeSentenceCase would
   therefore see \printfield commands rather than the field
   contents. If you want to convert all titles to sentence case,
   \MakeSentenceCase must be applied on the inner level, by the
   \printfield commands. To facilitate that, I've modified this and
   similar bibmacros slightly:

   \newbibmacro*{title}{%
     \ifthenelse{\iffieldundef{title}\AND\iffieldundef{subtitle}}
       {}
       {\printtext[title]{%
	  \printfield[titlecase]{title}%
	  \setunit{\subtitlepunct}%
	  \printfield[titlecase]{subtitle}}%
	\newunit}%
     \printfield{titleaddon}}
   
   Instead of 'noformat', the nested \printfield commands now use
   'titlecase' on the inner level. These formats are defined like
   this:

   \DeclareFieldFormat{titlecase}{#1}
   \DeclareFieldFormat{noformat}{#1}
   
   In other words, the default behavior remains unchanged but
   converting all titles to sentence case is as easy as saying:
   
   \DeclareFieldFormat{titlecase}{\MakeSentenceCase{#1}}
   
   You can apply additional formats on a higher level as usual:
   
   \DeclareFieldFormat{title}{\mkbibemph{#1}}
   \DeclareFieldFormat{title}{\mkbibquote{#1}}
   
   Just make sure that \MakeSentenceCase always hooks in on the
   innermost level such that it sees the raw field contents rather
   then data commands or other formatting commands.

RELEASE NOTES FOR VERSION 0.8

As usual, the full changelog is included in biblatex.pdf. What
follows are comments concerning changes 'under the hood' which are
not mentioned in the changelog, things which may break backwards
compatibility, improvements which may not be immediately obvious
from looking at the changelog, and new major features which pertain
to a set of entries in the changelog rather than a single one.

The changelog and the release notes include changes made in the
0.7a-g maintenance releases. If you have been tracking the
development closely, some things may sound familiar.

Note that there have been changes in the LaTeX<->BibTeX data
interface. You may want to delete all the old *.bbl and
<jobname>-blx.bib files before you start using the new version. In
fact biblatex should be able to handle the update on its own but
this will require an additional LaTeX+BibTeX cycle. In other words,
you need to go through one LaTeX/BibTeX/LaTeX cycle to update all
auxiliary files. You can ignore any warnings printed in this
process. After that, the output should stabilize as usual.

*  Hints for style developers

   I'm delighted to see that the first custom biblatex styles start
   showing up on CTAN.
   
   One general note to style authors: unless your style is only a
   minor modification of one of the standard styles which ship with
   biblatex, it's a good idea to make sure it's self-contained,
   i.e., that it doesn't employ \RequireCitationStyle and/or
   \RequireBibliographyStyle to load code from the default styles
   (but feel free to incorporate code by copying it to your style).

   While the core of biblatex is stable by now, there's still
   potential for modifications in the standard styles (see below for
   examples). Having said that, loading standard.bbx should be a
   safe thing to do and you can also rely on code in biblatex.def
   because this file is part of the core of biblatex.

*  New citation and bibliography styles

   I've added some new styles: 'authoryear-ibid' (an author-year
   style with an 'ibidem' feature), 'draft' (which is a, well, draft
   style), and 'reading' (a style for annotated bibliographies and
   personal reading lists with abstracts, annotations, etc.).

*  Overhaul of all citation and bibliography styles

   I've overhauled all styles such that they use LaTeX (=etoolbox)
   rather than plain TeX syntax for the boolean switches. E.g.,
   instead of:

   \newif\ifcbx@bool
   \cbx@booltrue
   \ifcbx@bool ...\else ...\fi
   
   they now use:
   
   \newbool{cbx:bool}
   \booltrue{cbx:bool}
   \ifbool{cbx:bool}{...}{...}
   
   This LaTeX frontend is provided by the etoolbox package.

*  Improved authoryear styles

   The authoryear styles have been improved such that they will
   consider the '(short)title' field if the author/editor is
   missing (or useauthor/useeditor=false is set). Note that these
   styles also consider the 'label' field, if available. In sum,
   the fallback chain in citations used to be:

   author -> editor -> label -> [issue warning]

   and the new chain works as follows: 

   author -> editor -> label -> shorttitle -> title

   The 'shorthand' field, if defined, always takes precedence over
   any other data.

*  Improved verbose-note styles

   The styles verbose-note and verbose-inote can now add a page
   number to the reference pointing to the initial, full citation.
   There is a style option named 'pageref' which enables this
   feature. The page reference is only printed if the initial
   citation is located on a different page or page spread (depending
   on the setting of the 'pagetracker' option).

*  Improved numeric styles

   The numeric styles now support citations referring to set
   members. See the note about reference sets below, the manual, and
   the examples for details.

*  Introducing reference sets

   This release introduces the concept of a reference or entry set.
   A reference set is a group of entries which are cited as a single
   reference and listed as a single item in the bibliography. I'm
   told that this is a matter of particular interest for users in
   physics, chemistry, and possibly some other fields. See the 'set'
   type, the 'entryset' field, and the special field 'entrysetcount'
   in the manual. Also see \entryset and the pointers in the
   changelog.

   If you are familiar with this concept, you may have used the
   mcite or the mciteplus package before. The mcite and mciteplus
   manuals call this type of reference a "collapsed citation". They
   also talk about "grouping citations". The biblatex manual calls
   it "reference set" and "entry set". It's the same thing but note
   that the approach is different. With mcite(plus), sets are
   defined as they are cited for the first time. Essentially, a
   citation like

   \cite{key1,*key2,*key3}
   
   defines a set consisting of three entries (key1, key2, key3). The
   first entry key serves as identifier of the entire set, which may
   subsequently be cited as \cite{key1}. With biblatex, you declare
   sets in the bib file using the @set type, the 'entryset' field,
   and 'entryset' reverse pointers in the child entries:

   @Set{set1,
     entryset = {key1,key2,key3},
     crossref = {key1},
   }
   @Article{key1,
     entryset = {set1},
     ...
   }
   @InCollection{key2,
     entryset = {set1},
     ...
   }
   @InProceedings{key3,
     entryset = {set1},
     ...
   }

   See the pointers in the changelog for further explanation.

*  Support for electronic publishing information (eprint)

   See the 'eprint' and 'eprinttype' fields and related pointers in
   the manual. Support for electronic publishing information has
   been added to all standard bibliography styles.

   This release comes with dedicated support for arXiv references
   which is usable out of the box and also serves as a model for
   adding support for other resources and online archives.

*  Modified 'labelyear' field

   The 'labelyear' field is no longer a string but an integer. The
   conversion of the integer to a string now takes places on the
   LaTeX side of the workflow, i.e., you'll usually have a
   formatting directive like this:

   \DeclareFieldFormat{labelyear}{\mknumalph{#1}}

   The \mknumalph command takes an integer in the range 1-702 as
   its argument and converts it to a string as used in author-year
   citations like "Jones 1995a". The format is:

   \mknumalph{1}   -> a
   \mknumalph{26}  -> z
   \mknumalph{27}  -> aa
   \mknumalph{702} -> zz

   There is also a 'maxlabelyear' counter which holds the highest
   number found in any 'labelyear' field. This may be useful if you
   want to print the 'labelyear' field as a number and pad it out
   with leading zeros.

*  Modified 'labelalpha' field

   There is a similar change concerning the 'labelalpha' field,
   which has been split up into 'labelalpha' and 'extraalpha'. With
   a label like "Jon95a", 'labelalpha' holds "Jon95" while the extra
   letter is handled by 'extraalpha'. 'extraalpha' is similar to
   'labelyear' in that it holds an integer. The conversion of the
   integer to a string takes places on the LaTeX side of the
   workflow, i.e., you'll usually have a formatting directive like
   this:

   \DeclareFieldFormat{extraalpha}{\mknumalph{#1}}

   There is also a 'maxextraalpha' counter which holds the highest
   number found in any 'extraalpha' field.

*  Modified/extended name hashes

   Starting with this release, there are two name hash fields,
   'namehash' and 'fullhash'. See the manual for details.
   Essentially, the difference is that the 'namehash' is derived
   from the visible 'labelname' list (subject to the 'maxnames' and
   'minnames' options) whereas 'fullhash' is always derived from the
   full list.

   Note that the original behavior of 'namehash' in previous
   releases used to be similar to the current behavior of
   'fullhash'. This has caused some confusion. The 'namehash' should
   now be in line with the expectations of most users and style
   authors. Those who actually want the original behavior may still
   use the 'fullhash' field.

*  'edition' field now more flexible/controllable by styles

   In previous biblatex releases, 'edition' was an integer-only
   field. The integer test was performed on the BibTeX side of the
   workflow, in biblatex.bst. Starting with this release, the test
   is performed on the LaTeX side, i.e., in the format definition:

   \DeclareFieldFormat{edition}{%
     \ifinteger{#1}
       {\mkbibordinal{#1}~\bibstring{edition}}
       {#1}%
   }
   
   As you can see, the edition is printed as "Nth edition" if the
   edition field holds an integer, and as a literal string if not.
   This means that you can now put things like "5th, revised and
   expanded edition" in the 'edition' field without having to resort
   to the 'note' field. It also implies that styles get full control
   over the 'edition' field.

   Style authors who adapt this formatting directive should note
   that ordinals are gender-specific in Romanic languages ('edition'
   is a feminine noun in French, Italian, and Spanish, hence the
   ordinal must be feminine as well). In addition to the default
   definition in biblatex.def, the 'edition' field is therefore
   adapted in french.lbx, italian.lbx, and spanish.lbx.

*  Support for bibliographic data in external TeX files

   This release add a \printfile command which is similar to
   \printtext but gets the text from an external file, i.e.,
   \printfile{file.tex} boils down to \printtext{\input{file.tex}}
   but does nothing if the file does not exist. The point of all
   this is that styles which print the fields 'abstract' and/or
   'annotation' may support an alternative way of adding abstracts
   or annotations to the bibliography. biblatex.def provides the
   following code for that:

   \newcommand*{\bibabstractprefix}{bibabstract-}
   \newcommand*{\bibannotationprefix}{bibannotation-}

   \newbibmacro*{annotation}{%
     \iffieldundef{annotation}
       {\printfile[annotation]{\bibannotationprefix
                               \thefield{entrykey}.tex}}%
       {\printfield{annotation}}}

   \newbibmacro*{abstract}{%
     \iffieldundef{abstract}
       {\printfile[abstract]{\bibabstractprefix
                             \thefield{entrykey}.tex}}%
       {\printfield{abstract}}}

   Instead of including the text in the bib file, it may now be
   stored in an external LaTeX file. For example, instead of saying

   @Article{key1,
      abstract = {This is an abstract of entry `key1'.}
      ...

   in the bib file, you create a file named 'bibabstract-key1.tex'
   and put the abstract in this file. The name of the external file
   must be the entry key prefixed with 'bibabstract-' or
   'bibannotation-', respectively. The 'reading' style makes use of
   this. Also note that, when using the reference code above, a field
   in the bib file takes precedence over external files. E.g., the
   'reading' style will not look for 'bibabstract-key1.tex' if the
   'key1' entry in the bib file has an 'abstract' field.

   Note that this feature needs to be enabled explicitly by setting
   the package option 'loadfiles'. The option is disabled by default
   for performance reasons. Using external files is strongly
   recommended if you have long abstracts or a lot of annotations
   since this may increase memory requirements significantly. See
   the changelog for more pointers.

*  Support for 'American-style' punctuation

   Biblatex finally supports 'American-style' punctuation, i.e.,
   certain punctuation marks placed after a closing quote can be
   moved inside the quotes automatically.

   See \DeclareQuotePunctuation and \mkbibquote in the manual. The
   'american' and 'canadian' localization modules enable this
   feature for periods and commas. See the hints in the manual for
   details.

*  \DeclareCapitalPunctuation replaces \(Enable|Disable)CapitalAfter

   The configuration commands \EnableCapitalAfter and
   \DisableCapitalAfter have been removed and are superseded by
   \DeclareQuotePunctuation.

   \DeclareQuotePunctuation works like \EnableCapitalAfter except
   that it takes a list of characters as its argument. The function
   of \DisableCapitalAfter is now implicit (all characters not
   included in the list will not trigger capitalization).
   
   The change affects the configuration interface only. The default
   user-level behavior of biblatex has not changed (capitalization
   after periods, exclamation marks, questions marks; with the
   '(n)german' and '(n)austrian' localization modules also after
   colons).

*  Hyphenation exceptions in bibliography strings

   This release adds an interface for definining hyphenation
   exceptions in lbx files and/or the document preamble. See
   \DeclareHyphenationExceptions and \DefineHyphenationExceptions
   in the manual.

*  Configurable number tests

   The number tests \ifnumeral and \ifnumerals, which are also used
   by \mkpageprefix and \mkpagetotal, may be adapted. See the
   commands \DeclareNumChars, \DeclareRangeChars, and
   \DeclareRangeCommands for details.

*  Added 'firstinits' package option

   By popular request, I've added a package option which switches
   all first names to initials. Use the \iffirstinits test to query
   its state. The test has already been incorporated into the
   formatting directives in biblatex.def. However, authors of custom
   styles using redefined name formatting directives may want to
   incorporate it into their style, too.

*  Added high-level '\bibpagespunct' macro

   \bibpagespunct is a high-level user macro similar to 
   \labelnamepunct and \subtitlepunct. As with the 'firstinits'
   package option, it's incorporated in the standard styles so you
   may want to do the same in custom styles.

*  Added some expert entry options

   See the entry options 'skipbib', 'skiplos', 'skiplab', and
   'dataonly' in the manual. 

   Note that 'skiplab' and 'dataonly' are intended for hacking only!

*  Improved KOMA-Script and Memoir support

   The headings 'bibliography' and 'shorthands' are responsive to
   the 'bibtotoc' and 'bibtotocnumbered' class options of the 'KOMA'
   classes now. See also \ifkomabibtotoc, \ifkomabibtotocnumbered,
   and \ifmemoirbibintoc.

*  Improved natbib compatibility style

   I've added some missing features to the natbib compatibility
   style. Most notably, citation aliasing is now possible. Note that
   this is intended for legacy files only. In newly created files,
   it is preferable to use biblatex's 'shorthand' field.

*  Spaces in file names

   Whether file names may or may not contain spaces generally
   depends on whether or not the underlying TeX binary supports
   that. This is beyond biblatex's control. If the underlying engine
   supports it, biblatex should not have any problems with spaces in
   a file name.

   However, neither traditional BibTeX nor bibtex8 seem to be
   capable of handling .bib files with spaces in their name. For
   .bib files which supply bibliographic data, there is no fix short
   of avoiding spaces in the file name. For the auxiliary .bib file
   automatically generated by biblatex this release adds a
   workaround. Spaces in the name of this file, which is only used
   internally, will be converted to underscores.

*  New language-specific csf files for bibtex8

   I've added some csf file for use with bibtex8 to the 'resources'
   subdirectory. They implement the proper sorting order for German,
   Danish, Norwegian, and Swedish in Latin 1, Latin 9, and Windows
   Ansi encoding.

   German users should note that even bibtex8 is not able to handle
   the German letter \ss properly. You may need to resort to fields
   like 'sortname' and 'sorttitle' even when using these csf files.

*  Extended manual

   In addition to the usual manual updates related to new features,
   I've also added some new material to the 'hints and caveats'
   section in the author guide of the manual.

RELEASE NOTES FOR VERSION 0.7

The changelog of this release is the longest list of changes in any
biblatex release so far. Skimming the full list in biblatex.pdf is
highly recommended. In the following, I will focus on changes 'under
the hood' which are not mentioned in the changelog, point out things
which may break backwards compatibility, and comment on some
improvements which may not be immediately obvious from looking at
the changelog because they pertain to a set of changes rather than a
single one.

*  User and author interface stable

   Starting with this release, the user and author interface of this
   package may be considered as stable. This means that I will try
   to refrain from making syntactically or functionally incompatible
   changes to the core package, unless I'm forced to do so because
   of a bug or because one of the most recently added features turns
   out to be really awkward. Essentially, citation and bibliography
   styles written for this version should work with biblatex 1.0
   with no or at most minor modifications.
   
   Please note that this is a development aim, but there is no
   guarantee. Also note that the definitions in biblatex.def (mainly
   the formatting directives) are not guaranteed to be stable at
   this point. In other words: if you use the default definitions,
   your style will inherit possible improvements in future versions.
   If you use modified formatting directives, you may need to update
   them to integrate new functions. That should be fairly easy,
   though.

*  CTAN staging area for biblatex styles

   There's a new location for contributed biblatex styles, it's:

   macros/latex/exptl/biblatex-contrib/
   
   Contributed styles should go into a subdirectory of the above
   location. E.g., a style called 'MLA' would go into:
   
   macros/latex/exptl/biblatex-contrib/mla/
   
   Note that the exptl/ subtree is intended for experimental or beta
   code. As soon as biblatex hits 1.0 and moves out of exptl/ to

   macros/latex/biblatex/
   
   there's be a corresponding biblatex-contrib. At this point, all
   contributed styles should be considered experimental because the
   core package is still in beta.

*  Removed biblatex.cbx and biblatex.bbx

   I've removed the files biblatex.cbx and biblatex.bbx from the
   distribution. The code formerly found in biblatex.cbx has been
   moved to biblatex.def. The code formerly found in biblatex.bbx
   has been split up. The most generic parts have been moved to
   biblatex.def. The less generic parts, which are closely tied to
   biblatex's standard bibliography style, are now part of
   standard.bbx.

   I've also rearranged some of the bibmacros formerly found in the
   above files and renamed some of them. If you have been using some
   of that code in a custom bibliography style, don't worry, all the
   code is still there, but parts of it may be arranged differently
   (e.g., a bibmacro may have been split up into two macros or two
   macro may have been merged into one) and the names of some
   bibmacros may have changed.

   Note that biblatex.def is always loaded, but standard.bbx is not.
   If you have been using code which is now part of standard.bbx in
   a custom bibliography style, you need to copy the relevant parts
   to your style.

*  New dependency on etoolbox.sty

   Some of the most generic parts of biblatex.sty have been moved to
   an independent package called 'etoolbox', which is now required.
   The package is available from CTAN and may also be useful for
   style authors.

*  Type-specific formatting directives

   All formatting directives may now be defined on a per-type basis;
   e.g.:
   
   \DeclareFieldFormat[article]{title}{\mkbibquote{#1}}
   
   would define the format of the 'title' field of @article entries
   only. The fallback mechanism for formats now works as follows.
   The command \printfield{title}, when used in an @article driver,
   would check for the following formats (in this order):
   
   title [article]
   title
   default
   
   This also applies to cases where a specific formatting directive
   is requested explicitly, i.e. \printfield[myformat]{title}, would
   check for these formats:

   myformat [article]
   myformat
   default

   The point is that you can override formats on a per-type basis
   without having to hack any drivers.

*  Truncation with 'and others' now supported by literal lists

   Truncating a list with 'and others' in the .bib file is now
   supported by both name lists and literal lists. This used to be
   specific to name lists. See \ifandothers and \ifmoreitems as well
   as \finallistdelim and \andmoredelim for details.

*  Introducing entry options and style-specific options

   This release introduces the concept of an entry option and a new
   'options' field. The 'useprefix' field has been replaced by an
   entry option. Instead of:
   
   useprefix = {true}
   
   you now use

   options = {useprefix=true}

   There are two more predefined entry options called 'useauthor' and
   'useeditor'. These options may still be used globally to set the
   default behavior.
   
   In addition to that, Bibliography and citation styles may define
   additional package and entry options. See \DeclareEntryOption and
   \DeclareBibliographyOption for details.

*  Introducing multicite commands

   This release introduces an entirely new class of citation
   commands called 'multicite' commands. The point of a multicite
   command is that its argument may be a fully qualified list of
   citations where each key has its own pre- and postnote. The syntax
   is straightforward. Instead of:
   
   \cite[See][55]{key1}; \cite[12]{key2}; \cite[93]{key3}
   
   you can now say:
   
   \cites[See][55]{key1}[12]{key2}[93]{key3}
   
   This is particularly useful with parenthetical citations and
   citations given in footnotes. It's also possible to assign a pre-
   and/or postnote to the entire list. These global notes are given
   in parentheses:

   \footcites(See)(and chapter 3)[55]{key1}[12]{key2}[93]{key3}
   
   See \cites, \parencites, \footcites, etc. in the manual. There is
   also an \autocites command (a multicite version of \autocite).

   Defining new multicite commands is very easy because they are
   based on the regular citation commands. See the documentation of
   \DeclareMultiCiteCommand for details.

*  Improved trackers and tracking control

   The 'trackers' provided by biblatex have been greatly improved
   and extended. This release also adds new trackers which
   correspond to the requirements of some common citation styles.
   See the package options 'pagetracker', 'citetracker',
   'ibidtracker', 'opcittracker', and 'loccittracker' for all the
   gory details.  Also see \ifciteibid, \ifopcit, and \ifloccit.
   It's possible to control tracking in the document, see
   \pagetrackertrue/false and \citetrackertrue/false for details.
   Note that text commands like \citetitle are now exluded from
   tracking by default. See also the point below.

*  Tracker reset support

   It's possible to reset trackers and citation styles and several
   different levels. For the built-in trackers, see the package
   option 'citereset' as well as the user-level command \citereset.
   For style-specific reset support, see \InitializeCitationStyle,
   \InitializeBibliographyStyle, \OnManualCitation, and the
   user-level command \mancite.

*  New entry types, more custom entry types

   This release adds the entry types 'periodical' and 'patent',
   which are fully supported by the standard styles. Several new
   custom types have also been added. The custom types are not
   supported by default, but they may be useful in custom styles.

*  New field/list subtype: key field/list

   I've introduced a new field/list data subtype called key
   field/key list. They may hold printable data or localization keys
   and work as follows: A test is performed to determine whether the
   value of the field is a known localization key. If so, the
   localized string is printed. If not, the value is printed as is.

*  Improved postnote handling, alternative pagination schemes

   The handling of page numbers in the 'pages' field and in the
   postnote argument to citation commands has been improved
   significantly. Biblatex does range detection now and recognizes
   Roman numerals as numbers. It also supports alternative
   'pagination' schemes (columns, line or verse numbers, etc.).

   See the fields 'pagination' and 'bookpagination' and related
   pointers in the changelog. See also \mkpageprefix, \ppspace,
   \pno, \ppno, \nopp, \psq, \psqq.

*  Support for unique names in citations

   This release introduces support for unique names in citations.
   Some author-year and author-title citation styles require
   unambiguous names in citations. For example, with entries such as

   John Smith 1995
   Edward Smith 1995
   
   citations would be rendered as
   
   J. Smith 1995
   E. Smith 1995
   
   rather than
   
   Smith 1995a
   Smith 1995b
   
   See the package option 'uniquename' and the special counter
   'uniquename' for details. Note that this feature is now enabled
   by default in the following styles: authoryear, authoryear-comp,
   authortitle-terse, authortitle-tcomp.

*  Improved support for numeric labels

   The 'defernums' package option addresses the problem of
   discontinuous numbering when using a numeric style in combination
   with bibliography filters. If this option is enabled, the numeric
   labels are assigned the first time an entry is printed in any
   bibliography. This is similar to the traditional algorithm used
   by LaTeX to assign numeric labels.

*  Alternative, non-inheriting cross-referencing mechanism

   This release introduces an alternative cross-referencing
   mechanism which does not inherit any data. It's useful in styles
   which format cross-referenced entries differently. See the
   description of the field 'xref' and related pointers in the
   manual.

*  Support for different encodings

   Biblatex is now capable of handling .bib files with an encoding
   which is different from the encoding of the .tex file. See the
   'bibencoding' package option for details.

*  More robust handling of citation keys

   This release adds some normalization code which deals with
   special characters in citation keys. A typical example are keys
   which contain an underscore. This should be much more robust now.

*  Renamed fields

   The 'journal' field has been renamed to 'journaltitle' but the
   old name is still supported as an alias. There's also a
   'journalsubtitle' field and fields for the title and the subtitle
   of a single issue. In other words, article entries now support a
   complete set of titles similar to inbook-like entry types:
   journaltitle/issuetitle/title essentially correspond to
   maintitle/booktitle/title.

   The field name 'id' turned out to be incompatible with JabRef.
   Apart from that, the name is a bit too generic for what this
   field holds anyway. It's called 'eid' now (electronic ID) and
   holds an article ID used by online journals or journals which are
   also available online.

*  'labelctitle' replaced by 'singletitle'

   The 'labelctitle' field has removed in favor of a different
   mechanism using a test called \ifsingletitle. The package option
   'labelctitle' has also been renamed to 'singletitle'.

*  Alphabetic label now configurable

   The alphabetic label provided in the 'labelalpha' field is now
   configurable to a certain extend. It is responsive to the global
   package options 'maxnames' and 'minnames'. It's also possible to
   influence the handling of truncated labels by redefining
   \labelalphaothers.

*  Name types

   This release introduces the concept of a name type. This is best
   explained by example. Suppose a book has a compiler rather than
   an editor. In previous versions of this package, there was no way
   to override the string 'editor' and 'edited by'. Starting with
   this release, you may specify 'compiler' as an editor type:

   @book{...,
     editor     = {...},
     editortype = {compiler},
   
   Supported editor types are 'editor' (the default) and 'compiler'.
   Note that this mechanism is hooked up to bibliography strings
   hence it may be extended. See the fields 'authortype',
   'editortype', and 'name[a-c]type' in the manual.

*  Consistent set of wrappers

   I've completed the generic wrapper commands provided by biblatex
   so that they form a complete set. Instead of using the solution
   in the left column of the following list, use the wrapper in the
   right column:

   (...)                   ->   \mkbibparens{...}
   [...]                   ->   \mkbibbrackets{...}
   \footnote{...}          ->   \mkbibfootnote{...}
   \textsuperscript{...}   ->   \mkbibsuperscript{...}
   \emph{...}              ->   \mkbibemph{...}
   \enquote{...}           ->   \mkbibquote{...}
   ``...''                 ->   \mkbibquote{...}

   The wrappers integrate much better with biblatex and they also
   provide additional features.

*  New citation styles, some styles renamed

   I've renamed all verbose citation styles from 'authortitle-*' to
   'verbose-*' and added some new styles. See the changelog for
   details.

*  natbib compatibility style

   To facilitate the move from natbib to biblatex, this release adds
   a special natbib compatibility style which maps natbib's core
   citation commands to equivalent biblatex commands. See the 
   package option 'natbib' for details.

*  All citation commands scan ahead for punctuation

   Starting with this release, all citation commands scan ahead for
   punctuation to avoid double punctuation at the end of a citation.
   This used to be a feature exclusive to \autocite. The 'autopunct'
   package option and the \DeclareAutoPunctuation command will now
   affect all citation commands, not only \autocite.

*  Support for font style adaptation of punctuation

   This release implements an alternative way of dealing with
   punctuation after a field printed in a different font (for
   example, a title printed in italics). The standard (La)TeX way of
   dealing with this problem is to add a small amount of space (the
   so-called italic correction) to avoid clashes between the final
   letter of a word in italics and the following (upright)
   punctuation mark.

   Biblatex is now capable of adapting the punctuation to the font
   of the preceeding field. See '\mkbibemph' and '\setpunctfont' for
   details. Note that this feature is experimental. It may very well
   have a few quirks. Also note that it is disabled by default. Use
   the 'punctfont' package option to enable it.

*  Extended language support: Norwegian, Danish

   This release comes with a norsk.lbx and a danish.lbx file. The
   translations have been contributed by Johannes Wilm.

RELEASE NOTES FOR VERSION 0.6

*  Style-independent citations

   This release introduces a special command for style-independent
   citations. The idea behind the \autocite command is to provide
   higher-level citation markup which makes global switching from
   inline citations to citations given in footnotes (or as
   superscripts) possible.

   The \autocite command is built on top of lower-level commands
   like \parencite and \footcite. The citation style provides an
   \autocite definition by way of \DeclareAutoCiteCommand. This
   definition may be activated by way of the 'autocite' package
   option. See the documentation of \autocite in the manual for
   further details.

*  Forcing capitalized name prefixes

   One thing that's been on my personal wishlist for some time is an
   equivalent to natbib's \Citet command. This release finally
   introduces additional citation commands which force capitalized
   name prefixes (provided that there is a prefix and that it is to
   be printed as part of the citation, i.e. the 'useprefix' option
   is enabled). The new citation commands are defined as follows in
   biblatex.cbx:

      \newcommand*{\Cite}{\bibsentence\cite}
      \newcommand*{\Textcite}{\bibsentence\textcite}
      \newcommand*{\Parencite}{\bibsentence\parencite}
      \newcommand*{\Footcite}{\bibsentence\footcite}
      \newcommand*{\Citeauthor}{\bibsentence\citeauthor}

   Here's how it works. Biblatex's punctuation tracker is based on
   TeX's space factor. All the \bibsentence command does is setting
   the space factor to a special sentinel value which is detected by
   \bibstring. The name formatting directive used by all citation
   commands now incorporates the new \ifcapital test. This test is
   true if the punctuation tracker would capitalize a bibliography
   string at this point. If the formatting directive detects the
   sentinel value, it will capitalize the name prefix (if
   applicable). A handy side-effect of this approach is that a
   possible 'prenote' argument is taken into account automatically.
   For example:

      '\cite{vandoren}'      prints 'van Doren 1995'
   
   and

      '\Cite{vandoren}'      prints 'Van Doren 1995'
   
   but

      '\Cite[See]{vandoren}' prints 'See van Doren 1995'.
   
   This also works in conjunction with bibliography strings. For
   example, if the citation style replaces the standard citation by
   something like '\bibstring{ibid}', then

      '\cite{vandoren}'      prints 'ibid.'
   
   but

      '\Cite{vandoren}'      prints 'Ibid.'.
   
   In other words, unless the citation style is doing something very
   unusual, there is no need to define \Cite et al. explicitly with
   \DeclareCiteCommand. The default definitions should work fine in
   all normal cases. If the style doesn't print any names but rather
   numerical or alphabetic citations then nothing is capitalized, so
   that's no problem either.

*  Custom filters for \printbibliography

   See \defbibfilter and the 'filter' option of \printbibliography
   in the manual.

*  Support for unsorted bibliographies

   See the package option 'sorting=none'.

*  Automatic truncation of literal lists

   This release introduces the package options 'maxitems' and
   'minitems' which are similar to 'maxnames' and 'minnames' but
   affect literal lists. There are also corresponding (local)
   options for \printbibliography, \printshorthands, and so on.

*  Improved support for corporate authors and editors
   
   There are two new name lists ('shortauthor' and 'shorteditor')
   which may be helpful when dealing with corporate authors and
   editors. Basically, 'author' and 'editor' are used in the
   bibliography whereas 'shortauthor' and 'shorteditor' are used in
   citations. The point is that you can give a short form of the name
   for use in citations. For example:
   
   @Type{key,
    author = {{National Aeronautics and Space Administration (NASA)}},
    shortauthor = {NASA},
    ...
   }
   
   This will print "NASA" in citations but "National Aeronautics and
   Space Administration (NASA)" in the bibliography. Note that
   'shortauthor' and 'shorteditor' are name lists, not literal
   fields. This means that corporate names must be wrapped in an
   additional pair of curly braces.

*  Improved support for articles: journal series and electronic ID

   Journal series and article IDs are now catered for by the
   standard styles (entry type 'article'). Note that the 'articleid'
   field has been renamed to 'eid'. See the annotated bibliography
   in the 'examples' directory for examples.

*  DOI support

   The 'doi' field is now catered for by the standard styles
   (complete with hyperlinks). See the annotated bibliography in the
   'examples' directory for examples.

*  Fine-grained control of title formatting

   Previous versions of this package provided two formatting
   directives for the title field: 'title' (for entry types like
   'book', 'collection', etc.) and 'titlein' (for 'article',
   'inbook', 'incollection', etc.). I've enhanced this scheme such
   that the format of the title may be defined on a per-type basis.
   In biblatex.def you'll find the following directives for the
   bibliography:

   \DeclareFieldFormat{title:book}{\emph{#1}\isdot}
   \DeclareFieldFormat{title:inbook}{\enquote{#1}\midsentence}
   ...

   as well as dedicated directives for the titles in citations:

   \DeclareFieldFormat{citetitle:book}{\emph{#1}\isdot}
   \DeclareFieldFormat{citetitle:inbook}{\enquote{#1}\midsentence}
   ...

   and in the list of shorthands:

   \DeclareFieldFormat{lostitle:book}{\emph{#1}\isdot}
   \DeclareFieldFormat{lostitle:inbook}{\enquote{#1}\midsentence}
   ...

   I've updated all generic definitions in biblatex.cbx,
   biblatex.bbx, and standard.bbx accordingly. Note that the
   'title:type' directives are used for both the 'title' and the
   'subtitle' field (which are wrapped in a \printtext command
   controlling the formatting, see the 'title+stitle' bibmacro in
   biblatex.bbx for an example of how it works). In other words:
   redefining the 'subtitle' field formatting directive has no
   effect.

*  Rearranged localization keys

   I've rearranged and extended the localization keys quite a bit.
   This also implies several changes in biblatex.bbx.

*  Support for KOMA-Script and Memoir

   The default definitions of the bibliography headings (as defined
   in biblatex.def) are now automatically adapted for the
   KOMA-Script classes and the Memoir class.

*  Improved Spanish support

   The Spanish localization module now handles the Spanish word
   'and' properly ('y' or 'e', depending on the context).

*  Italian support

   This release comes with a new italian.lbx file. The translations
   have been contributed by Enrico Gregorio.

RELEASE NOTES FOR VERSION 0.5

   My original plan for 0.5 was to make minor changes only and focus
   on fixing bugs. Since only a few issues turned up (memory issues
   in biblatex.bst), I ended up implementing new things from the
   wishlist. Please note that the wishlist is closed until after
   version 1.0.

*  New author-title citation styles

   I've added some verbose author-title styles which may be of
   interest for users in the humanities. The styles are called
   'authortitle-verb' and 'authortitle-cverb'. The 'traditional
   style has been renamed to 'authortitle-trad' and there is an
   additional new style called 'authortitle-strad'. The styles
   support shorthands and are fully hyperlinked. The links point to
   the first, verbose citation instead of the bibliography so these
   styles may be used without a bibliography, if desired.

   I've also added a matching 'dummy' bibliography style for every
   citation style. The dummy style will simply load one of the more
   generic backend styles. For example, the 'authortitle-verb' style
   loads the 'authortitle' style. The point is that you may simply
   use the 'style' package option instead of 'citestyle' plus
   'bibstyle'.

*  Conditional special fields, optimizations in biblatex.bst

   The special fields 'labelalpha', 'labelctitle', 'labelnumber',
   and 'labelyear' are conditional now. If a style requires one of
   those fields, it needs to request them by setting the
   corresponding package option. E.g., all numeric styles include
   the line

      \ExecuteBibliographyOptions{labelnumber}

   This will instruct biblatex to provide the 'labelnumber' field
   which is undefined by default now. Making these fields
   conditional allows for some memory-related optimizations in
   biblatex.bst.

*  Author-level support for full citations

   This version adds an author-level command which makes it possible
   to execute a bibliography driver in a citation command. See
   '\rundriver' in the manual.

*  Page tracker

   I've implemented a page tracker. The page tracker records the
   real page on which citations, entries in the bibliography, and
   entries in the list of shorthands end up. It can work on a per
   page basis or per double page spread. The page tracker is
   disabled by default. See the package option 'pagetracker' and the
   commands '\iffirstonpage' and '\ifsamepage' in the manual.

*  Unique indentification of reference instances

   Every 'instance' of a reference is now uniquely identified by a
   value of the 'citecount' counter. In other words: this counter is
   incremented for every key processed by any citation command, for
   every item in the bibliography and for every item in the list of
   shorthands. This is required by the page tracker but it's also
   useful if you need to generate a unique anchor name for
   hyperlinks.

*  Extended support for hyperlinks

   I've renamed '\bibhyperlink' to '\bibhyperref' and added generic
   '\bibhypertarget' and '\bibhyperlink' commands. '\bibhyperref'
   creates a link from a citation to the corresponding item in the
   bibliography. '\bibhypertarget' and '\bibhyperlink' are more
   generic and correspond to the '\hypertarget' and '\hyperlink'
   commands of the hyperref package. The point here is that you
   don't need an explicit '\ifhyperref' check. If hyperlinks are
   disabled, these wrappers will simply pass on their text argument.

   As an additional benefit, they also provide better anchor
   placement. The anchors created by '\hypertarget' seem to be
   located at the baseline, i.e. if you click on a link you get the
   impression that your PDF viewer jumps to the line of text just
   below the one you mean.

*  More name lists

   See the 'annotator', 'commentator', 'introduction', 'foreword',
   and 'afterword' lists in the annotated example.

*  Special fields for indexing

   See the 'indextitle' and 'indexsorttitle' fields in the annotated
   example and the indexing examples.

*  Spanish support

   This release comes with a preliminary spanish.lbx file. The
   translations have been contributed by Ignacio Fern\'andez
   Galv\'an. Some advanced features are missing from this file
   because they are still under scrutiny but it should nevertheless
   be perfectly funtional.

*  Examples

   I've added a biblatex showcase with some example files in the
   'examples' subdirectory. There's an example for every citation
   style and some generic examples demonstrating multiple
   bibliographies, split bibliographies, indexing, and so on.

RELEASE NOTES FOR VERSION 0.4

*  Wishlist closed

   As of this release, the wishlist is closed until after version
   1.0. There are already more wishes in the pipeline than I can
   implement in time for 1.0...

RELEASE NOTES FOR VERSION 0.3

   The changelog for this release is shockingly long so I'll just
   point out the general developments behind the individual entries
   in the changelog. See the full changelog for all the gory
   details. Note that styles written for version 0.2 will not work
   with 0.3.

*  New data type: literal list

   This release introduces a new data type for literal lists. The
   point of a literal list is that the list is split up at the 'and'
   but the individual items are not dissected further. This is
   intended for fields such as 'location' and 'publisher' since they
   may contain a list of items but these items are not personal
   names.

   I have modified the macro names of the entire data interface for
   lists to match the change. All macros with the term 'list' in
   their name are renamed such that 'list' is replaced by 'name(s)'.
   For example, \printlist is now \printnames, \DeclareListFormat is
   \DeclareNameFormat, and so on. The old names are used for new
   macros which deal with literal lists. The complete list of
   affected macros is given in the changelog. Note that the names of
   the macros which dump the unformatted data have changes as well
   (\biblist -> \thename; \thelist (new); \bibfield -> \thefield).

   These changes will require updates to all custom citation and
   bibliography styles. However, the changes are only a matter of
   some search & replace commands. I'm sorry about the inconvenience
   but I think the new naming scheme is better than having macro
   names like \printliterallist. I've decided to go for it now since
   biblatex is still in beta.

   As part of these changes, the data type of the following fields
   has been changed from 'literal field' to 'literal list':
   location, origlocation, publisher, institution, organization.

*  Support for hyperref

   Hyperref support is available now. This means that citations may
   be transformed into links pointing to the bibliography. All
   standard citation styles support that out of the box, custom
   styles need to use the new hyperref interface. It works like
   this: the anchor (i.e. the target of the link) is set
   automatically by biblatex hence the bibliography style does not
   need to do anything special. The citation style is responsible
   for marking the link area. This is just a matter of passing the
   relevant part of the citation to a special macro or formatting
   directive doing the low-level work. See the changelog for
   pointers to the relevant sections of the manual.

*  Support for back references

   I've added support for back references. The page numbers are
   provided in the 'pageref' field which uses the new literal list
   data type. Printing them is just a matter of \printlist{pageref}.
   Note that you need to enable the 'backref' package option to get
   any back reference data. There is also a bibmacro called
   'pageref' in biblatex.bbx which adds a label. Bibliography styles
   should print the list via \usebibmacro{pageref}. There are two
   related list formatting directives in biblatex.def. The default
   directive just prints the list as is, using a comma as a
   separator. The directive 'pageref:comp' prints a sequence of more
   than two consecutive pages as a range. The easiest way to try it
   out is: \DeclareListAlias{pageref}{pageref:comp}.

   The references are restricted to page numbers, back references to
   sections are not supported. This is a deliberate decision. I
   haven't bothered implementing back references to sections because
   LaTeX's referencing mechanism is notoriously unreliable in this
   respect. It essentially provides two pieces of information: the
   formatted value of the last counter incremented by way of
   \refstepcounter and the page number. The latter value is usually
   unambiguous but the former could refer to just about everything.
   The implications of this become obvious when you're using the
   backref package with its 'ref' option (rather than 'pageref') and
   put \cite commands in footnotes. The bibliography will then
   include something like 'sections 1, 2, 3', but these numbers do
   not refer to sections, these are footnote numbers!

   Getting back references to sections right would require
   reimplementing LaTeX's entire referencing mechanism (or
   interfacing with a package doing that) and I don't want to get
   into that business just yet. I may look into support for the zref
   package later, but that's something for a post-1.0 release of
   biblatex.

*  Rearranged author-title citation styles

   I've rearranged the author-title citation styles because most
   people seem to expect the behavior of the old 'authortitle-verb'
   style from the plain 'authortitle' style. So 'authortitle-verb'
   is the plain 'authortitle' style now. The style formerly known as
   'authortitle' has been renamed to 'authortitle-terse',
   'authortitle-comp' is 'authortitle-cterse' now. I've also added a
   new 'authortitle-comp' style to round off this subset of styles.

*  Handling of thebibliography/theshorthands streamlined

   The handling of thebibliography and theshorthands as well as some
   related facilities has been overhauled and streamlined. See the
   changelog and section 4.2.2 of the manual for details.

*  Handling of 'and others' simplified

   The way biblatex handles the string 'and others', which is used
   in bib files to truncate a list of names, has been overhauled and
   simplified. Essentially, I have removed 'moreauthor' and similar
   fields since biblatex handles this internally now. All style
   authors need to do is use \ifandothers and/or \ifmorenames. See
   the manual for details.

*  Improved \addtocategory

   \addtocategory now cycles its arguments through the aux file.
   This means that it may be used in the preamble and anywhere in
   the document body, even after \printbibliography.

RELEASE NOTES FOR VERSION 0.2

*  Internal changes in bibliography styles

   I have rearranged all bibliography styles. It should be more
   obvious now how many code is shared by the bibliography styles,
   namely all 'driver' code. The shared code has been moved to
   standard.bbx, so think of that as the biblatex standard style.
   This code is used by numeric.bbx, alphabetic.bbx,
   authortitle.bbx, and authoryear.bbx; standard.bbx is not meant to
   be used stand-alone. A few bibmacros have also been moved to
   biblatex.bbx.

*  Internal changes in citation styles

   There is only one change affecting the citation styles. The
   'postnote' bibmacro in biblatex.cbx now automatically inserts a
   prefix like 'p.' or 'pp.' where applicable.

*  Optional argument for \printtext

   \printtext takes an optional argument now. The point is that you
   can use \printtext as a formatting hook. The advantage of this
   approach is that \printtext integrates with the punctuation
   tracking. I have revised biblatex.bbx to use this approach where
   applicable.

*  New fields

   There are several new fields, some of which are supported by the
   standard styles. This has lead to further changes in
   biblatex.bbx, including changes to the names of existing
   bibmacros.
