A Look at HTML+ and the HTML 3.0 Draft

In parallel to HTML's development by Tim Berners-Lee and Dan Connolly, Dave Raggett, one other prominent guru in the development of HTML was working on his own expanded and extended version of HTML which he called "HTML+." I have not included the versions of HTML+ (nor the HTML 3.0 Draft which is also largely Dave Raggett's and draws much from his work with HTML+) since these works contain much in the way of what might properly be considered proposed elements ("tags") and attributes, some of which would later come to be implemented in HTML 4.01 (and 4.0), and beyond, and the remainder of which are largely not implemented by any contemporary user agents (the way that, for example, <XMP> continues to be implemented in virtually every modern user agent despite the extreme obsolescence of this element) and as such have gone the way of proposed elements as <TYPEWRITER> and <KEYWORDS>.

Because Dave Raggett's work takes a fundamentally different tack than that of the others, his versions of HTML+ simply cannot be placed in any sequential order with the versions of HTML as listed in my main Lost Tags of HTML file, and these significant early versions of HTML with such a significant number of properly "lost" tags and attributes would have to be addressed at least somewhere. Like in the main Lost Tags of HTML file, this file contains lists of comparitive versions of the HTML, showing differences between the versions. However, in this case, a slightly different set of standards for comparison have been used. For one thing, in this file the disused tags are not cumulative among the versions (as they are in the main Lost Tags file). Comparisons are made strictly to the previous version of HTML+ (if any), and to the version of HTML which is current at the time of the release of the particular version of HTML+ or HTML 3.0.

In these quick listings of the tags (elements) and attributes comparing the versions of HTML+ to each other and do the versions of HTML current at the time of each, I have utilized the following color scheme to represent the various categories of the state of each tag and of each attribute, so as to make comparison between the various versions easier at a glance. These are the colors used:

And as with the main sequence of HTML I have provided working test DTDs. Again, this small note here, regarding details of the DTDs that I have prepared or adapted for these unofficial versions of HTML, explains in detail how they can be used for experimentation.

"HTML 1.q" The oldest surviving version of HTML+

The oldest surviving description of HTML+ dates from November 1993, but hints contained in the description point to a previous version of HTML+ (now lost) which appears to have dated from a period even prior to the publication of "HTML 1.k" in that most of the basic HTML+ description files contain the <HEADER> element - thus dating their origin to that period from November 26, 1992 to (at least) May 24, 1993, or else (at least as likely), Dave Raggett was still using the version of the NeXT HTML editor that Tim Berners-Lee had used for that crucial period, instead of some more recent or current editor. Seeing the character entity reference errors that HTML 1.k introduced in this version of HTML+ does suggest that he may have used that published draft as his starting point, thus dating his first efforts after that time. These same files were then hand-modified as shown by the inclusion of the HTML+ element <HR> as well as the descriptions of the later HTML+ features. This earlier version of HTML+ had some differences from the November 1993 version of HTML+ (which I call here "HTML 1.q" to be consistent with my versioning system introduced in my main Lost Tags of HTML file). These differences can be gleaned by a couple means within the description files, first by descriptions of tags that are not in the DTD (such as the conditional text tags, <ONLINE> and <PRINTED>) but described as though they have always been around, and not appearing in any subsequent version of HTML+, second by descriptions of tags or attributes, as seen in one of the Math examples, an <RD> element, meant to serve as a right delimiter (for some reason the pipe ("|") character was not meant to expand like the other brackets and parenthesis, "(" ")" "{" "}" "[" and "]", would when encompassing a taller expression in a math formula, so various means were being used for implementing a stretchable pipe delimiter character, useful for such things as matrix determinates, such as &ldet; and &rdet; as well as <RD> (and by implication <LD> as well). These character entities and elements are not recognized by any surviving form of HTML+, nor by the HTML 3.0 Draft. There is also a reference to their having once been an <EMBED> element which would have enabled one to embed graphic information of some sort into the HTML file itself, and that this element had already been abandonded by the time of this November draft. By mistake, several other elements were intended to be included in this version of HTML+ but minor errors in the DTD make the inclusion of these other elements impossible, namely <NEXTID>, <ISINDEX>, <BLOCKQUOTE>, <ADDRESS>, <TR>, and <RENDER>.

"HTML 1.q" - (the oldest surviving release of HTML+) as compared to "HTML 1.k" contains the following differences:

Comments
Document Level Tags
Document Head Tags
Block Level Body Tags
Raw Text Tags
Hyper Link Tags
User Input Tags
Listing Tags
Dictionary Tags
Mathematic Equations Tags
Inline Data Type Tags
Inline Formatting Tags
Inserted Object Tags
Table Tags

I have adapted this DTD to a format usable with the W3C validating engine and it can be invoked by affixing the following declaration to a file:

<!DOCTYPE HTMLPLUS SYSTEM "http://www.the-pope.com/html1.q.dtd">

A special version, created to correct some errors in the DTD (and which adds <ISINDEX>, <NEXTID N>, <TR ID>, <RENDER TAG STYLE>, <ONLINE>, <PRINTED>, and the basic HTML character entities &amp;, &lt;, &gt;, and &quot;) to make it more consistent with what seems to have been intended can also be invoked by affixing the following declaration to a file:

<!DOCTYPE HTMLPLUS SYSTEM "http://www.the-pope.com/html1.qc.dtd">

"HTML 1.s" The second and last surviving version of HTML+

The other surviving version of HTML+ (what I call "HTML 1.s") dates from April 1994, and is currently pointed to as the "Final version of HTML+" and so it seems to be as none later has been found or ever hinted at. There are some hints in the comments to the DTD (and a couple other places in the description) that there may have been some intermediate version of HTML+ between that of November 1993 and this version of April 1994, but these hints are real skimpy and provide little information as to what that might have actually looked like in any detail.

For the second and final surviving version of HTML+ I have deviated here in this listing from my usual policy of going by the defaults as to what might be included or excluded depending upon the use of various SGML selector mechanisms. The current "HTML 1.m" standard employed several of these selectors with which to select minimal, obsolete, and prescriptive versions of the language, as well as to be able to select or deselect certain individual elements, KEY, U, and NEXTID. I used the defaults for these, which gave the largest and most advanced form of the language and excluded only the KEY and U elements. Following this precedent, the final version of HTML+ also employs quite a number of such SGML selectors, this time being for inclusion or exclusion of math, tables, figures, emph (emphasis), forms, obsolete (such as <BLOCKQUOTE>, not <XMP> which is already completely gone!), and icons. The default would be to exclude all of these features, making this version of HTML+ quite dull, so for this listing I have included all seven of these categories. It would be like using the following SGML declaration:

<!DOCTYPE htmlplus [
        <!ENTITY % HTML.math "INCLUDE">
        <!ENTITY % HTML.tables "INCLUDE">
        <!ENTITY % HTML.figures "INCLUDE">
        <!ENTITY % HTML.emph "INCLUDE">
        <!ENTITY % HTML.forms "INCLUDE">
        <!ENTITY % HTML.obsolete "INCLUDE">
        <!ENTITY % HTML.icons "INCLUDE">
]>

Taking that exception into account, "HTML 1.s" - (the latest surviving release of HTML+) as compared to "HTML 1.m" and also to "HTML 1.q" contains the following differences:

Comments
Document Level Tags
Document Head Tags
Block Level Body Tags
Raw Text Tags
Hyper Link Tags
User Input Tags
Listing Tags
Dictionary Tags
Mathematic Equations Tags
Inline Data Type Tags
Inline Formatting Tags
Inserted Object Tags
Table Tags

"HTML 3.0" A Pass at the Futire - Building on HTML+ and HTML 2.0

HTML 2.0 officially introduced a stict versioning system to the versions of HTML that were being published, and with that Dave Raggett's work changed its name from HTML+ to a working proposal (draft) for the next version of HTML, namely 3.0. Even well before the final official text for HTML 2.0 was published in September 1995 he and others were already working on HTML 3.0 with the clear intention that it would be the successor to HTML 2.0. Using his final version of HTML+ as a starting point he incorporated ideas from many others, interested parties, industry, academia, and even ideas that the Berners-Lee/Connolly group were tossing on the table, thus capturing in quite some detail the current ideas for HTML's future. The final draft of this version of HTML was last updated in March 1995, before HTML 2.0 was officially finalized, but as it was intended to be a proper superset of what HTML 2.0 provides and largely suceeded at that despite coming so much earlier (only known exceptions would be the URN attribute of <A> and the various non-SGML plaintext and literal elements), I have compared it to HTML 2.0 instead of the previous "HTML1.m" which was still the most current published version at the time.

The March 1995 draft of HTML 3.0 as compared to HTML 2.0 and also to "HTML 1.s" contains the following differences:

Comments
Document Level Tags
Document Head Tags
Block Level Body Tags
Raw Text Tags
Hyper Link Tags
User Input Tags
Listing Tags
Dictionary Tags
Mathematic Equations Tags
Inline Data Type Tags
Inline Formatting Tags
Inserted Object Tags
Table Tags

Looking at the final actual accepted version of HTML 3.2 by comparison, one has to admit that HTML 3.2 was far less sophisticated than the proposed 3.0, and indeed many of the presentational elements and attributes that would soon after have to be depreciated came not from Dave Raggett's work, but from the much less ambitious and more cautious approach taken by a Web Consortium which had at about this time also found itself suddenly bereft of IETF support. So much of the infrastructure needed to support style sheets that 3.0 provided was gutted in favor of simplistic presentational features never intended for HTML but which the industry was coming up with in the absence of any clear support for HTML 3.0, let alone any consistent or universal standard for any sort of style sheets. Only with the coming of HTML 4.0 (and 4.01) would this condition finally be rectified, and now there was a whole host of these now-useless presentational features which exist in HTML 4 but only as depreciated elements and attributes.

This file, "htmlplus.html," is HTML 4.01 Transitional compliant.

Next Level Up