<chapter id="change-list">
<title>DocBook Changes Made and Planned</title>
<para>This chapter records the changes made to the several most recent releases
of the DTD and the changes planned for future releases. See <xref linkend="upgrading">
for information on upgrading to new versions of DocBook.</para>
<sect1>
<title>Changes Made from V2.1 to V2.2.1</title>
<para>From V2.1 to V2.2.1 most changes were minor, although a few were backwards
incompatible. (As noted in <xref linkend="rev-policy">, future backward incompatible
changes are now made only in major DTD revisions.) The following changes in
that release were backwards incompatible:<itemizedlist><listitem><para>The
elements Character, Charset, Font, and Glyph were removed, as were EventStructure,
EventType, Mask, and ProtocolRequest.</para>
</listitem><listitem><para>The FAXTILE notation was removed.</para>
</listitem><listitem><para>IndexTerms were disallowed from appearing in Highlights.
</para>
</listitem><listitem><para>Footnote was excluded from itself (beware, all
you folks who were footnoting your footnotes!).</para>
</listitem><listitem><para>The Mark attribute on FootnoteRef (which is to
be renamed Label in V3.0) was changed to <literal>#IMPLIED</literal> from <literal>
#CONREF</literal>.</para>
</listitem><listitem><para>The Width attribute was removed from ScreenShot.
</para>
</listitem><listitem><para>Logo was removed from the Class attribute on ProductName
and Trademark.</para>
</listitem><listitem><para>The Linkend attribute was removed from ToCFront.
</para>
</listitem><listitem><para>The Linkend and PageNum attributes formerly found
on the components of ToC were moved to the new ToCEntry element.</para>
</listitem></itemizedlist>The following were the significant backwards-<emphasis>
compatible</emphasis> changes in V2.2.1:<itemizedlist><listitem><para>A notation
declaration was added for CGM-BINARY.</para>
</listitem><listitem><para>ProductName and ProductNumber were made available
within <sgmltag class="paramentity">docinfo.content.gp</sgmltag>, and thus
as inline markup in most places.</para>
</listitem><listitem><para>Optional was made available as inline markup.</para>
</listitem><listitem><para>Subtitle was expanded to be able to contain inline
markup.</para>
</listitem><listitem><para>All the link elements were expanded to be able
to contain inline objects, and were given common attributes. XRef was given
common attributes in addition to Linkend and Endterm.</para>
</listitem><listitem><para>LegalNotice was expanded to be able to contain
many additional block-oriented elements.</para>
</listitem><listitem><para>In Revision, AuthorInitials and RevRemark were
made optional.</para>
</listitem><listitem><para>Application, CiteTitle, and Quote were expanded
to be able to contain a wider range of inline elements.</para>
</listitem><listitem><para>Subscript and Superscript were expanded to be able
to contain Emphasis.</para>
</listitem><listitem><para>GlossSee and GlossSeeAlso were expanded to be able
to contain inline elements.</para>
</listitem><listitem><para>PrimaryIE, SecondaryIE, TertiaryIE, SeeIE, and
SeeAlsoIE were expanded to be able to contain inline elements, and their Linkends
attributes were changed to <literal>#IMPLIED</literal> from <literal>#REQUIRED
</literal>.</para>
</listitem><listitem><para>Contrib was created and included in the content
models of Author, Editor, and OtherCredit, as a place to store information
about the contributions of these parties to the document in question. It contains
plain text.</para>
</listitem><listitem><para>GlossList was created: it's a set of GlossEntries
and may be used anywhere lists are allowed.</para>
</listitem><listitem><para>InformalExample was created, as a parallel to other
informal block-oriented elements. It's just like Example, but without a title.
</para>
</listitem><listitem><para>ToCentry was created and added to all the components
of ToC, to hold the actual entries in a Table of Contents. It may contain
plain text and in-line elements.</para>
</listitem><listitem><para>Address was expanded to be able to contain the
new OtherAddr for modes of addressing not now covered by other elements.</para>
</listitem><listitem><para>Block-oriented elements were allowed to occur between
SubSteps.</para>
</listitem><listitem><para>Para was expanded to be able to contain InlineGraphic
and InlineEquation.</para>
</listitem><listitem><para>Entry (a table cell) was expanded to be able to
contain elements from <sgmltag class="paramentity">code.example.gp</sgmltag>
(LiteralLayout, ProgramListing, Screen, and ScreenShot).</para>
</listitem><listitem><para>Glossary was changed to be able to close with an
optional Bibliography.</para>
</listitem><listitem><para>A RefNameDiv was allowed to contain more than one
RefClass.</para>
</listitem><listitem><para>Editor and OtherCredit were added as allowable
contents in AuthorGroup.</para>
</listitem><listitem><para>A new Scope attribute was added to IndexTerm with
the allowed values All, Global, and Local (the default is <literal>#IMPLIED
</literal>), to meet the case of IndexTerms that should be used to generate, <foreignphrase>
e.g.</foreignphrase>, an index for a Book (Local) but not an index for a Set
(Global) or vice versa; the value All should be used when the IndexTerm should
appear in all generated indices. You should define your implementation's default
accordingly.</para>
</listitem><listitem><para>Also, a new Zone attribute, with <literal>#IMPLIED
</literal> <literal>IDREFS</literal> values, was added to IndexTerm. This
attribute is to be used when it is desired to attach indexing information
to SGML elements rather than, or in addition to, placing IndexTerms in the
flow of the text so as to be sure page references are generated correctly.
Zone could be used to attach keywords to sections and paragraphs, for example,
facilitating the assembly of nonstructural views of the indexed text.</para>
</listitem><listitem><para>The RenderAs attribute default for Sect1&ndash;5
and Bridgehead was changed to <literal>#IMPLIED</literal> (from the same value
as the element's generic identifier, or Other, in the case of Bridgehead&mdash;there
might be some backward incompatibility for BridgeHead).</para>
</listitem><listitem><para>A SrcCredit attribute was added to Graphic, InlineGraphic,
and LoTEntry for storage of information crediting the source of an image.
</para>
</listitem><listitem><para>Many more inline elements were given common attributes.
</para>
</listitem><listitem><para>The ID attributes on Footnote and BeginPage were
changed to <literal>#IMPLIED</literal> from <literal>#REQUIRED</literal>.
</para>
</listitem><listitem><para>The PageNum attribute on Anchor was changed to <literal>
#IMPLIED</literal> from <literal>#REQUIRED</literal>.</para>
</listitem><listitem><para>Part was added as a defined value for the Pubwork
attribute on CiteTitle.</para>
</listitem><listitem><para>GlossEntry was given a SortAs attribute, like IndexEntry.
</para>
</listitem><listitem><para>The Subject attribute on GlossDef was changed from <literal>
NMTOKENS</literal> to <literal>CDATA</literal>.</para>
</listitem></itemizedlist></para>
</sect1>
<sect1>
<title>Changes Made in V2.3</title>
<para>The markup model of DocBook changed in the following ways in V2.3 as
a result of the dramatic changes to the DTD's architecture (which are not
detailed here):<itemizedlist><listitem><para>ModeSpec was allowed in <sgmltag
class="paramentity">setinfo.char.mix</sgmltag> contexts. Its omission was
discovered during the restructuring.</para>
</listitem><listitem><para>Because the CALS-based table model now uses a parameterization
design based on work done by SGML Open, several table customizations in V2.2.1
were no longer possible in V2.3:<orderedlist><listitem><para>The common attributes
were disallowed from ColSpec and SpanSpec.</para>
</listitem><listitem><para>The default value of the VAlign attribute on both
THead and TFoot was changed to <literal>#IMPLIED</literal>. Previously, the
default was Bottom for THead and Top for TFoot.</para>
</listitem></orderedlist>These changes were backwards incompatible, but regression
testing showed no problem with large collections of legacy documents; that
is, no organizations appeared to be using these markup features.</para>
</listitem></itemizedlist></para>
<para></para>
</sect1>
<sect1>
<title>Changes Made in V2.4 and V2.4.1</title>
<para>V2.4 has had a patch release, V2.4.1; you should consider all of the
following changes together.</para>
<bridgehead>V2.4 Changes</bridgehead>
<para>The following global and multiple-module changes were made:<itemizedlist>
<listitem><para>All the modules associated with DocBook have been changed
to use the new version number in their comment headers and formal public identifiers,
as appropriate. The <filename>docbook.cat</filename> file contains the corresponding
updated FPIs.</para>
</listitem><listitem><para>In <filename>docbook.dtd</filename>, <filename>
dbhier.mod</filename>, <filename>dbpool.mod</filename>, and <filename>docbook.cat
</filename>, the entity declarations and corresponding catalog entries for
the <quote>rdb</quote> placeholder entities were removed. If users need these,
they should declare them and manage their catalog entries themselves, and
should not use <quote>Davenport</quote> as the owner identifier for these
entities.</para>
</listitem><listitem><para>In <filename>dbhier.mod</filename> and <filename>
dbpool.mod</filename>, the order of attributes was changed to put the unique
(element-specific) ones first, then the semi-common ones, then the common
ones. The placeholder for local attribute extensions is still at the end.
</para>
</listitem><listitem><para>In <filename>dbhier.mod</filename> and <filename>
dbpool.mod</filename>, many elements now have an <literal>OMITTAG</literal>
scheme of <literal>"-&nbsp;O"</literal> instead of <literal>"-&nbsp;-"</literal>.
In general, series of similar elements don't require an end-tag.</para>
</listitem><listitem><para>In <filename>dbhier.mod</filename> and <filename>
dbpool.mod</filename>, the comments documenting the common attributes now
provide information on the assumed default values if an attribute value is
not supplied.</para>
</listitem><listitem><para>In <filename>dbhier.mod</filename> and <filename>
dbpool.mod</filename>, comments were added for planned backwards-incompatible
changes in V4.0, and the comments for the V3.0 changes have been made to stand
out more. (Search in the DTD files for the string <quote>FUTURE USE</quote>,
or for <quote>FUTURE USE (V<replaceable>n</replaceable>.0)</quote> with the
relevant version number supplied.)</para>
</listitem></itemizedlist></para>
<para>The following changes were made to <filename>docbook.dtd</filename>:<itemizedlist>
<listitem><para>Added the BMP, WMF, PCX, and WPG (WordPerfect Graphic) notations
and added to notation.class correspondingly.</para>
</listitem><listitem><para>Fixed the owner identifier in the entity declaration
for the <quote>grk4</quote> ISO entity set (Alternative Greek Symbols); it
was <quote>ISO 9573-13:1991</quote>, but has been changed to <quote>ISO 8879:1986
</quote>.</para>
</listitem></itemizedlist></para>
<para>The following changes were made to <filename>dbhier.dtd</filename>:<itemizedlist>
<listitem><para>Modularity and parameterization changes:<itemizedlist><listitem>
<para>The name of the sect1.content entity has been changed to bookcomponent.content
to reflect its purpose more accurately.</para>
</listitem><listitem><para>Added a second redeclaration placeholder to allow
redeclaration of the bookcomponent.content entity while letting it retain
its reference to divcomponent.mix.</para>
</listitem></itemizedlist></para>
</listitem><listitem><para>New elements:<itemizedlist><listitem><para>Added
a new element, SimpleSect, to sect1.content (used in Appendix, Chapter, Preface,
PartIntro, and Article) and directly to the content of Sect2 through Sect
5.</para>
</listitem></itemizedlist></para>
</listitem><listitem><para>Content model changes:<itemizedlist><listitem>
<para>Added SimpleList to indexdivcomponent.mix (used in introductions to
indexes).</para>
</listitem><listitem><para>The content of Book now allows a mixture of zero
or more Glossary, Bibliography, and Preface elements before the main content,
and a mixture of zero or more Glossary and Bibliography elements after the
main content.</para>
</listitem><listitem><para>BookInfo and DocInfo now allow zero or more graphics,
to describe the document's content graphically.</para>
</listitem><listitem><para>The content of ToC and Index is now allowed to
be entirely empty (with no subelements supplied).</para>
</listitem><listitem><para>The ToCfront, ToCentry, and LoTentry elements now
use the new para.char.mix mixture instead of the defunct inline.char.mix mixture
(see dbpool.mod).</para>
</listitem><listitem><para>Chapter and Appendix now allow an optional ToCchap
before their main content.</para>
</listitem><listitem><para>The PrimaryIE, SecondaryIE, TertiaryIE, SeeIE,
and SeeAlsoIE elements now contain the new ndxterm.char.mix mixture instead
of the defunct inline.char.mix mixture (see dbpool.mod).</para>
</listitem><listitem><para>RefMiscInfo now contains the docinfo.char.mix mixture
instead of just <literal>#PCDATA</literal>.</para>
</listitem><listitem><para>RefDescriptor now contains the refname.char.mix
mixture instead of just <literal>#PCDATA</literal>.</para>
</listitem><listitem><para>The content of RefSynopsisDiv has been expanded
to match that of RefSect1.</para>
</listitem></itemizedlist></para>
</listitem></itemizedlist></para>
<para>The following changes were made to <filename>dbpool.dtd</filename>:<itemizedlist>
<listitem><para>Modularity and parameterization:<itemizedlist><listitem><para>
Changed the name of the entity containing the common attributes for use on
graphics (along with its corresponding <quote>local</quote> entity) from graphic.attrib
to graphics.attrib, and changed its corresponding references in Graphic and
InlineGraphic accordingly.</para>
</listitem></itemizedlist></para>
</listitem><listitem><para>New elements:<itemizedlist><listitem><para>Added
element declarations and corresponding module entities for GUIButton, GUIIcon,
GUILabel, GUIMenu, GUIMenuItem, and GUISubmenu and added these elements to
cptr.char.class. These are the elements bursted out from Interface. Put in
a FUTURE USE comment on Interface for removing its Class attribute in V4.0.
Added a declaration for the Accel element and added it to the content models
of Interface and the GUI* elements.</para>
</listitem><listitem><para>Added element declarations and corresponding module
entities for KeyCombo and MouseButton and added these elements to cptr.char.class.
</para>
</listitem><listitem><para>Added element declarations and corresponding module
entities (and an aggregate module entity) for MenuChoice and Shortcut and
added MenuChoice to cptr.char.class. Added a <quote>key action</quote> attribute
entity for use in Shortcut and KeyCombo.</para>
</listitem><listitem><para>Added the new Area, AreaSet, AreaSpec, Callout,
CalloutList, GraphicCO, ProgramListingCO, and ScreenCO elements to support
callouts. Corresponding parameter entities for their marked-section modules
(as well as the aggregate module entities for AreaSpec content and CalloutList
content) have been added as well. CalloutList has been added to the list.class
class. ProgramListingCO and ScreenCO have been added to the linespecific.class
class. GraphicCO has been added to the informal.class class and to the content
of ScreenShot.</para>
</listitem><listitem><para>Added a CO element declaration and corresponding
module entity, and added CO to the declarations of the program listing and
screen content model entities.</para>
</listitem><listitem><para>Added the new Attribution element (along with its
marked-section module entity); it appears in the content models of BlockQuote
and Epigraph.</para>
</listitem><listitem><para>Added the new FuncPrototype element (along with
its marked-section module entity); it appears in the content model of FuncSynopsis.
</para>
</listitem><listitem><para>Added the new Phrase element (along with its marked-section
module entity); it appears in the word.class class.</para>
</listitem><listitem><para>Added the new SBR element (along with its marked-section
module entity); it appears in the content of the CmdSynopsis, Arg, and Group
elements.</para>
</listitem></itemizedlist></para>
</listitem><listitem><para>Content model changes:<itemizedlist><listitem>
<para>Moved the SGMLTag and Markup elements to the cptr.char.class entity
to allow these elements in the same locations as other technical inlines (such
as RefName).</para>
</listitem><listitem><para>Added <literal>#PCDATA</literal> to the Address
content model, added the Format (linespecific) attribute to it, and added
the element to informal.class. (Note that legacy addresses may not work correctly
under new linespecific Address assumption!)</para>
</listitem><listitem><para>Added Email to cptr.char.class and moved its declarations
accordingly.</para>
</listitem><listitem><para>Added the formal.class class to the admon.mix and
glossdef.mix mixtures.</para>
</listitem><listitem><para>Added the formal.class class to the para.mix mixture,
and removed it by means of an SGML exclusion from the content models of Footnote,
Highlights, Example, and LegalNotice, where paragraphs (now normally allowing
formal-object content) can occur. (It was not excluded from Procedure, MsgExplan,
MsgText, or BlockQuote, as originally planned, because these elements can
legitimately contain formal objects as part of their component.mix content.)
</para>
</listitem><listitem><para>Changed the entire mixture scheme for inlines.
Removed the inline.char.mix, synop.char.mix, ssscript.char.mix, and phrase.char.mix
mixtures, and added the title.char.mix, ndxterm.char.mix, smallcptr.char.mix,
word.char.mix, and docinfo.char.mix mixtures. The following tables show approximately
what has changed. (Individual elements that use different mixtures are mentioned
below.)</para>
<para>New: <programlisting>                    #PCD xref word link cptr base dnfo othr inob (synop)
para.char.mix         X    X    X    X    X    X    X    X    X     X
title.char.mix        X    X    X    X    X    X    X    X    X
ndxterm.char.mix      X    X    X    X    X    X    X    X    a
cptr.char.mix         X              X    X    X         X    a
smallcptr.char.mix    X                   b                   a
word.char.mix         X         c    X         X         X    a
docinfo.char.mix      X         c         b              X    a
  
a. Just InlineGraphic; no InlineEquation.
b. Just Replaceable; no other computer terms.
c. Just Emphasis and Trademark; no other word elements.</programlisting></para>
<para>Old: <programlisting>[X = all, . = some] #PCD xref word link cptr base dnfo othr inob (synop)
para.char.mix         X    X    X    X    X    X    X    X    X     X
inline.char.mix       X    X    X    X    X    X    X    X
synop.char.mix        X    X    X    X    X    X    X    X
cptr.char.mix         X              X    X    X         X
ssscript.char.mix     X         .    X    .    X         X
phrase.char.mix       X              X         X         X</programlisting></para>
</listitem><listitem><para>Changed the programlisting.content entity (used
in ProgramListing), the screen.content entity (used in Screen), and the Comment,
Seg, Member, Term, LiteralLayout, Synopsis, ProductName, Application, RefEntryTitle,
CiteTitle, and Quote elements to use the para.char.mix mixture instead of
the defunct inline.char.mix mixture.</para>
</listitem><listitem><para>Changed the Title, TitleAbbrev, SubTitle, BridgeHead,
and SegTitle elements to contain the new title.char.mix mixture instead of
the defunct inline.char.mix mixture.</para>
</listitem><listitem><para>Changed the BiblioMisc, MsgAud, ScreenInfo, Citation,
Emphasis, ForeignPhrase, and LineAnnotation elements to contain the para.char.mix
mixture instead of just <literal>#PCDATA</literal>.</para>
</listitem><listitem><para>Changed the tblexpt entity for exclusions to table
content to use the formal.class class (which includes Table) instead of just
Table.</para>
</listitem><listitem><para>Changed the Ackno, Street, POB, Postcode, City,
State, Country, Phone, Fax, Email, OtherAddr, ShortAffil, JobTitle, OrgDiv,
ArtPageNums, CollabName, AuthorInitials, ConfDates, ConfTitle, ConfNum, ConfSponsor,
ContractNum, ContractSponsor, Year, Holder, CorpAuthor, CorpName, Date, Edition,
ISBN, ISSN, InvPartNumber, IssueNum, ModeSpec, OrgName, PageNums, Contrib,
FirstName, Honorific, Lineage, OtherName, Surname, ProductNumber, PubDate,
PublisherName, PubsNumber, ReleaseInfo, RevNumber, RevRemark, SeriesVolNums,
and VolumeNum elements to use the new docinfo.char.mix mixture instead of
just <literal>#PCDATA</literal>.</para>
</listitem><listitem><para>Changed the MsgLevel, MsgOrig, ClassName, ErrorName,
ErrorType, KeyCode, KeySym, MediaLabel, ReturnValue, StructField, StructName,
Symbol, Token, Type, Markup, and SGMLTag elements to use the new smallcptr.char.mix
mixture instead of just <literal>#PCDATA</literal>.</para>
</listitem><listitem><para>Changed the Replaceable element to use a series
of class entities directly, plus newly containing the Optional and InlineGraphic
elements, instead of using the defunct phrase.char.mix mixture.</para>
</listitem><listitem><para>Changed the Abbrev,Acronym, ManvolNum, FirstTerm,
and WordAsWord elements to use the new word.char.mix mixture instead of just <literal>
#PCDATA</literal>.</para>
</listitem><listitem><para>Changed the Subscript and Superscript elements
to use a series of class entities directly, plus newly containing the Replaceable,
Symbol, and InlineGraphic elements, instead of using the defunct ssscript.char.mix
mixture.</para>
</listitem><listitem><para>Changed the Trademark element to use a series of
class entities directly, plus newly containing the InlineGraphic element,
instead of using the cptr.char.mix mixture.</para>
</listitem><listitem><para>Changed the Primary, Secondary, Tertiary, See,
and SeeAlso elements to use the new ndxterm.char.mix mixture instead of the
defunct inline.char.mix mixture.</para>
</listitem></itemizedlist></para>
</listitem><listitem><para>New and changed attributes:<itemizedlist><listitem>
<para>Fixed a bug in Anchor attribute list that allowed it not to have an
ID. This was an oversight in the V2.3 modularization.</para>
</listitem><listitem><para>Added a Spacing attribute on ItemizedList and OrderedList.
</para>
</listitem><listitem><para>Added a Path attribute to Filename element.</para>
</listitem><listitem><para>Added a new common RevisionFlag attribute.</para>
</listitem><listitem><para>Added the new common <quote>effectivity</quote>
attributes OS, Arch, Vendor, UserLevel, and Revision. Collected the effectivity
attributes in their own entity and reorganized that section of entities appropriately.
</para>
</listitem><listitem><para>Added a TermLength attribute to the VariableList
element.</para>
</listitem><listitem><para>Added a Class attribute to the Filename element.
</para>
</listitem><listitem><para>Added a Class attribute to the Symbol element.
</para>
</listitem><listitem><para>Added a Linkend attribute to the FirstTerm element.
</para>
</listitem><listitem><para>Added Linkend and BaseForm attributes to the GlossTerm
element.</para>
</listitem><listitem><para>Added Class attribute values to the SGMLTag element.
</para>
</listitem></itemizedlist></para>
</listitem></itemizedlist></para>
<bridgehead>V2.4.1 Changes</bridgehead>
<para>Following are the changes made to DocBook V2.4 to create V2.4.1:<itemizedlist>
<listitem><para>Substantive changes in <filename>docbook.dtd</filename>:</para>
<para>Three notation declarations (<literal>EQN</literal>, <literal>PIC</literal>,
and <literal>TBL</literal>) have been changed to have null system IDs because
the original identifiers were formed as approximate public IDs, not system
IDs (and because they contained an ampersand, they were not valid FPIs either).
We are investigating whether there are proper formal public IDs for these
notations.</para>
</listitem><listitem><para>Substantive changes in <filename>dbpool.mod</filename>:<itemizedlist>
<listitem><para>The vendor.attrib entity containing the Vendor effectivity
attribute (new in V2.4) has been added to effectivity.attrib in order to make
the attribute available for use.</para>
</listitem><listitem><para>The MenuChoice element (new in V2.4) has been added
to cptr.char.class in order to make the element available for use.</para>
</listitem><listitem><para>The content model of Shortcut (used inside MenuChoice)
was changed to match that of the new KeyCombo element, as its original design
intended.</para>
</listitem><listitem><para>The Option element was added to the content model
of Group in CmdSynopsis to bring the model up to what the documentation (correctly)
describes.</para>
</listitem></itemizedlist></para>
</listitem><listitem><para>Editorial changes:</para>
<para>The copyright statment has been updated to help ensure continued free
access to DocBook files by users.</para>
</listitem><listitem><para>Spelling in comments was fixed, and comments explaining
some attributes were clarified.</para>
</listitem></itemizedlist></para>
</sect1>
<sect1>
<title>Changes Planned for V3.0</title>
<para>We will make the following backwards-incompatible changes in V3.0: <itemizedlist>
<listitem><para>The SpanEnd attribute on IndexTerm will be renamed StartRef
to indicate that it points to the content of the IndexTerm being spanned.
A Class attribute will be added to indicate whether the IndexTerm element
is the beginning point of a span (StartOfRange), the ending point (EndOfRange),
or a standalone index term (Singular).</para>
</listitem><listitem><para>FootnoteRef will be declared <literal>EMPTY</literal>
and its Mark attribute renamed Label.</para>
</listitem><listitem><para>RevHistory will require at least one Revision.
</para>
</listitem></itemizedlist></para>
</sect1>
<sect1>
<title>Changes Planned for V4.0 to Date</title>
<para>We expect to make at least the following backwards-incompatible changes
in V4.0; more such changes may be planned in the coming months:<itemizedlist>
<listitem><para>Articles <emphasis>may</emphasis> be removed from Books and
put into their own top-level DTD hierarchy.</para>
</listitem><listitem><para>Most elements currently containing the cptr.char.mix
entity will have their content models restricted to <literal>#PCDATA</literal>
and a much smaller set of elements.</para>
</listitem><listitem><para>Epigraph will be disallowed from appearing in BlockQuote.
</para>
</listitem><listitem><para>The OptMult and ReqMult values for the Choice attribute
on Group will be removed; it is expected that the Rep attribute value Repeat
will be used instead.</para>
</listitem><listitem><para>The content model group in FuncSynopsis starting
with FuncDef will not be repeatable; it is expected that the FuncPrototype
element will be used for repeating prototype blocks instead.</para>
</listitem><listitem><para>Interface will no longer have a Class attribute;
it is expected that one of the more specific elements with <quote>GUI</quote>
as the root of their names, or the Interface element with a Role attribute
value, will be used instead.</para>
</listitem></itemizedlist></para>
</sect1>
</chapter>