     International Organisation for Standardisation
       International Electrotechnical Commission

               ISO/IEC JTC1/SC18/WG8

Title: Committee Draft proposal for the new HTML work item
Source: David Abrahamson, Ireland, david@tcd.ie and 
        Roger Price, France, rprice@cs.uml.edu
Project: new
Status: Proposed Committee Draft
Action: For consideration at the next SC18/WG8 meeting.
Date: 1996-xx-yy
Distribution: WG8 and liaisons
References: none

Introduction
------------

This contribution proposes a committee draft as a basis for developing an
ISO/IEC International Standard for the HyperText Markup Language
(HTML).

Committee Draft availability and contents
-----------------------------------------

The committee draft is available at
ftp://dmsl.cs.uml.edu/pub/strawman/strawman.tar.gz

When un`tar'ed, the contribution produces a subdirectory "strawman"
with the following files:

1) ./strawman/iso-html.sgml  ./strawman/standard.dtd 

A markup of the proposed committee draft of the proposed standard. It
uses Anders Berglund's DTD for ISO text.  Sorry, we dont have a
formatted version.

2) ./strawman/iso-html.dcl  ./strawman/iso-html.ent  
   ./strawman/iso-html.dtd

The SGML declaration, entity set, and the elements and attributes of
the proposed committee draft.

3) ./strawman/SDOcorba.dtd  ./strawman/SDOstyle.ent  
   ./strawman/SDOstyle.dtd

Examples of the SGML techniques that other standards development
organisations (SDO) might use to define standard extensions to the
ISO-HTML core language.  They are not intended as contributions to any
standards development organisation: they serve only to assist
understanding of the core language by illustrating possible uses of
the interfaces provided.

4) ./strawman/sample1.iso-html  ./strawman/sample2.iso-html
   ./strawman/sample1.esis      ./strawman/sample2.esis

An example of a simple `pure' ISO-HTML document, and an example of a
complex document which makes use of the "addon" interfaces to add the
inline style tags, and a sketch of an interface to OMG Corba objects.
The element structure information set files show what all conforming
parsers should understand from the markup.

5) Some other files that are useful if you use James Clark's nsgmls
SGML parser.

Differences between W3C HTML 3.2 and this committee draft
---------------------------------------------------------

Since the W3C HTML specification is still "work in progress" and not
suitable for reference, and since the committee draft is still under
development, this list is necessarily unstable - however it should
give an idea of the extent of the differences.

a) The committee draft assumes that the W3C's HTML 3.2 draft specification
includes the IETF HTML internationalisation features.  The differences
between the HTML internationalisation specification and this committee draft
are:

   1) N/A (The CLASS attribute in now included.)

   2) The internationalisation specification provides the ICADD fixed
   attributes to support easy transformation to their DTD.  Since these
   facilities are not included in W3C's HTML 3.2 draft, they are not
   included in this committee draft.  They could be added through the "addon"
   interfaces.

   3) The W3C's HTML 3.2 includes the elements <U>, <STRIKE>, <BIG>,
   <SMALL>, <SUB> and <SUP> in the parameter entity %font;.  These
   elements are not present in %font; in the internationalisation
   specification.  Consequently <SUB> and <SUP> have the content model
   (%text;)* rather than (#PCDATA) as called for by the
   internationalisation specification.

   4) The elements <DFN> and <Q> have been added to the parameter
   entity %phrase;.  <DFN> is included in %phrase; by the W3C's HTML 3.2
   and <Q>, introduced by the internationalisation specification, has
   been added to the parameter entity %phrase; by this committee draft 
   as a clarification of its role in the DTD.

   5) 
 
b) The committee draft conforms more strictly to the requirements of
ISO 8879:1986 subclause 15.5 "Documentation Requirements".

c) The committee draft does not offer the facilities classified as
deprecated by W3C's HTML 3.2 draft: The <XMP>, <LISTING> <PLAINTEXT>
elements and the horizontal tab character;

d) The start tag and end tag omission allowed by W3C's HTML 3.2 draft
for the document element <HTML> and the major elements <HEAD> and
<BODY> are not allowed by this committee draft.

e) The elements <DIR> and <MENU> are omitted since in practice they
are sugared syntax for the <UL> element.

f) The <CENTER> element is omitted since it is a special case of the
<DIV> element.

g) The start tag omission allowed by W3C's HTML 3.2 draft for
non-empty <TEXTFLOW> elements is not allowed by this committee draft.
<TEXTFLOW> is a required element which may be empty, and start tag
omission is not allowed by ISO 8879:1986 for empty required
elements. [Goldfarb 7.3.1.1 p.310 line 6]

h) Parameter entity references in the DTD are systematically
terminated by the refc `;'.  [See Goldfarb 9.4.5 p.352 for a
discussion of the conditions for omission of the refc.]

i) All uses of the ISO owner identifier for the SGML standard are now
"ISO 8879:1986".  [See Goldfarb 10.2.1.1 p.383]

j) The committee draft does not provide direct support for the <FONT>
and <BASEFONT> elements.  If such facilities are required, they may be
defined by use of the body.extend parameter entity interface. An
example of the use of the body.extend; interface is given in file
sample2.iso-html.

k) This committee draft defines a <NOP> element and a NOP attribute
for use by the DTD to ensure syntactically correct content models for
unused interfaces.  This element and attribute are not intended for
use in documents.

l) The inline style attributes for the <BODY>, <BR> and <HR> elements
which were not provided by IETF RFC 1866 but which were added by W3C's
HTML 3.2 draft have been removed from the committee draft.  They may
be added if needed by an application of this specification as addons,
as shown in the sample document.

m) The use of the `;' character as the separator for name/value pairs
in the form-urlencoded media type, encouraged but not mandated by IETF
RFC 1866 is required by this committee draft.

n) The correct nesting of the sections identified by the <H1> through
<H6> elements which was recommended by IETF RFC 1866 is required by
this committee draft.

o) In the <IMG> element, the ALT attribute, considered vital by W3C's
HTML 3.2 draft but optional, becomes mandatory, and the attributes
WIDTH, HEIGHT, HSPACE, VSPACE and BORDER not present in IETF HTML 2.0
but introduced by W3C's HTML 3.2 are omitted.  They may be added if
required using the "img.addon" interface.

p) The ACCEPT attribute of the <INPUT> element, proposed by RFC 1867
and mentioned in W3C's HTML 3.2 draft although not included in the
W3C's HTML 3.2 draft DTD, is not included in this committee draft.

q) The inline style attribute for the <INPUT> element specified by
IETF RFC 1866 and extended by W3C's HTML 3.2 draft has been removed
from the committee draft.  They may be added if needed by an
application of this specification as an addon.

r) The "place holder" elements <SCRIPT> and <STYLE> are replaced by
the head.extend interface technique which allows the addition of such
facilities when they are defined.

s) The committee draft defines an aligning element as an element
carrying an ALIGN attribute which specifies horizontal placement
relative to the current left and right margins.  The aligning elements
give priority to the innermost when nested.  Note that for historical
reasons, not all elements which carry the ALIGN attribute are aligning
elements. E.g. <CAPTION> is not an aligning element.

t) The term "character set" used by W3C's HTML 3.2 draft in the
specification of the <TEXTAREA> element has been replaced by the term
"coded character set" as defined by IETF RFC 1866.

u) The URN and METHODS attributes of the <A> element, included in IETF
HTML 2.0 and the internationalisation specification, but omitted by
W3C's HTML 3.2 have been omitted from this committee draft.  They may
be included if needed using the "addon" technique.

v) The NOSHADE, SIZE and WIDTH attributes of the <HR> element have
been omitted.  They may be added through use of a style sheet, or by
use of the hr.addon interface.

w) In the W3C HTML 3.2 specification the ALIGN attribute of the
<TABLE> element appears to give precedence to the aligning effect of
an outer enclosing <DIV> element.  Such outer precedence is not
supported in this specification.

x) The TYPE attribute defined by W3C HTML 3.2 for the LI, OL and UL
elements to provide a suggested presentation style is not included in
this committee draft.  It may be added through use of the addon
interfaces of those elements.

Acknowledgements
----------------

This proposal has benefited considerably from comment by IETF and W3C
experts in the iso-html mailing list.  The editors wish to express
their thanks to all those who participated in the discussion.
 
End of file
