[Archive copy mirrored from: http://www.ornl.gov/sgml/wg8/document/1931.htm, July 15, 1997]

ISO/IEC JTC 1/SC 18/WG8 N1931


Document Processing and Related Communication --

Document Description and Processing Languages

TITLE: Report on W3C HTML 4.0/ISO-HTML Harmonization Meeting
SOURCE: Roger Price
ACTION: For information
DATE: July 15th 1997
DISTRIBUTION: WG8 and Liaisons


The final Committee Draft of ISO 15445 ISO-HTML is currently under review, and balloting will end on August 15th 1997. If the text receives the approval of the ISO/IEC JTC1/SC18 National Body (NB) members, the editors will proceed with the disposition of comments and the preparation of the final text for the International Standard, which should be ready for publication before the end of 1997.

Concurrently the World Wide Web Consortium is preparing the specification of their HTML 4.0 Proposed Recommendation. There are differences between ISO-HTML and HTML 4.0 which could lead to confusion about the "correct" definition of the language.

In order to resolve this confusion, the ISO-HTML editors David Abrahamson and Roger Price met as authorised with the W3C HTML Architect Dave Raggett in Dublin from July 13th to July 15th to address the question of the harmonization of the two specifications for the HTML Language. Dave Raggett participated as an Invited Expert.

It is a pleasure to report that this meeting was very successful, and has succeeded in showing:

The participants reviewed in considerable detail the two HTML specifications and prepared a list of minor changes which your editors recommend WG8 adopt as a prerequisite to the review of comments on the final CD balloting.

This report provides a summary of the technical conclusions of the meeting, and a detailed technical report will be prepared separately.

ISO-HTML documents conform to W3C HTML 4.0

We have established that it is possible, with only minor changes to the ISO-HTML specification, to ensure that all ISO-HTML documents conform to W3C HTML 4.0. The minor changes were typically adjustments to sets of attributes and content models. No substantial changes were needed to major components of ISO-HTML which would require review of the specification by the NBs. Example of the changes are given later in this report.

We have not added to the structures of W3C HTML 4.0. This met the requirements of WG8 1927 "Recommendations of the Barcelona Meeting" recommendation 8.

Inline style

W3C HTML 4.0 provides a <STYLE> element and systematically uses a STYLE attribute to allow authors to add style specifications. At present ISO-HTML does not support inline style specifications in documents.

To harmonize these two positions, we recognized that there are two major classes of language user: those who need inline style mechanisms in the language, and those who require a language with a full separation of content and style - i.e. no inline style mechanisms in the language.

In order to satisfy these apparently conflicting requirements, we proposed adopting the W3C <STYLE> element and STYLE attribute in a supplementary specification to be placed in a normative annex. The mechanisms required to incorporate the extensions already exist in ISO-HTML. We will define a formal public identifier for "ISO-HTML plus Inline Style" which will facilitate the use of the extension by user agents. The user agents will not be required to support the document type declaration subset.

We propose removing presentation attributes such as ALIGN, VALIGN and COMPACT from all elements.


Following expert advice, we propose excluding IMAGE as a form input type. This part of the W3C specification is not yet sufficiently mature for inclusion in the ISO-HTML specification. The feature is subject to accessibility problems for non-graphical user agents, and is not sufficiently mature for adoption in ISO-HTML.


Following expert advice, we propose to drop the <ISINDEX> element from ISO-HTML.


We reviewed the W3C HTML 4.0 specification of tables in detail. This is an extension of the earlier 3.2 specification. We propose adopting the extended 4.0 specification which is a superset of the current ISO-HTML text, however, on expert advice, we propose to omit the presentation attributes.


The meeting confirmed the decision to include the <OBJECT> element in ISO-HTML.


The meeting confirmed the decision to exclude the <APPLET> element from ISO-HTML. We do not propose to add this element until advised to do so by the W3C. However, users may add the element using the "addon" technique within the document type declaration subset.


The ISO-HTML definition of the <IMG> element should be extended to include the ALTSRC attribute required for improved accessibility by people with disabilities. This attribute provides a URL pointing to an extended description.


We reviewed in detail the current W3C specification of scripting and noted a number of areas were the techniques are still too immature for inclusion in an ISO specification. We do not propose to include any scripting facilities until we are advised by the W3C that scripting has become sufficiently stable and comprehensive.


The meeting confirmed the decision to include the internationalization facilities defined by RFC2070 in ISO-HTML. The ISO-HTML specification should not repeat the character entity sets defined in W3C HTML 4.0, but should refer to them using a formal public identifier.

Inserted and deleted text

We propose to include the <INS> and the <DEL> elements in ISO-HTML. However we do not propose to adopt the DATETIME attribute until the specification has stabilized.


We propose to add the <ACRONYM> element to ISO-HTML.

Improved accessibility

We propose to adopt the ACCESSKEY attributes which provide for the use of single key characters to facilitate navigation, and the TABINDEX attribute which allows an author to specify the tabbing order.

We propose to adopt the use of the <LABEL>, <FIELDSET> and <LEGEND> elements which improve the structuring of forms and the user access to the form contents.


We propose to remove the REV attribute from the <LINK> element. This attribute offers no facility that cannot be obtained by other means and is the subject of confusion.


We have much pleasure in thanking Dave Raggett for his considerable assistance and invaluable help in this work.

Roger Price