Re: Suggestion for HTML Text Format Elements

Amanda Walker (
Wed, 5 Jul 95 12:43:20 EDT

> My fear is that content providers coming in
> with WYSIWYG DTP skills will be inclined to say "Hey, with Acrobat
> output I can use the same tools I had been and get a document that
> looks just the way I want it to and now the Web supports it, so now
> I don't need to learn HTML to be a good Web citizen."

I like this better than, "well, HTML won't do what I want so I'll just
make a big GIF image map."

> And the
> resulting documents will lose all their internal structure and will
> be much less useful (for searching, inclusion, and use by other
> tools) than they would have been if they had been appropriately marked
> up.

One of Acrobat's strengths (over, say, raw PostScript) is that an Acrobat
document does not, in fact, lose all of its internal structure. It's still
searchable, has internal and external links, headings, etc. And it's about as
usefully searchable as an HTML version (perhaps more so since Adobe actually
has supported searching and cataloging tools on store shelves now).

> ---
> | I'm a great fan of SGML as a means for doing markup--it's
> | great. However, not all problems are solved by marked up text.
> ---
> No, but it solves the vast bulk of problems for the kinds of
> conventional documents most people are creating.

Look at the Web. Most of it isn't text. Most people aren't creating
"conventional documents," they're creating ads and data sheets, or personal
scrapbooks. If people were publishing articles, manuals, etc., that would be
one thing, but by and large they're not (and not coincidentally, the folks
that are doing so are *not* the people pushing for visual tools).

> You cannot over-emphasize the advantage of providing a means for authors
> to assign semantic information to the elements of the documents
> they create.

Certainly; I'm a great fan of semantic markup (I was a Scribe weenie in the
DEC-20 days :)). However, much of the content people want to put on the Web
isn't being created by authors, and doesn't consist of "documents" in the
traditional sense. Much of it is being created by artists, who are creating
interactive artwork.

SGML is from the ground up a mechanism for associating semantic information
with a stream of text. At this task it excels, and is the best tool for that
purpose which mankind has yet devised.

However, not all content is a stream of text with a semantic structure
associated with it. Not all content, for that matter, is intended to be
available in a revisable form (some, in fact, will not be published if the
only choice is an easily revisable format). I think this is one of the
reasons that the New York Times chose Acrobat over HTML for their WWW
offering, for example. HTML certainly had the expressive power, but they
quite intentionally did not *want* to make the entire structure freely

> I'm just afraid that, just at the point where it looked like we might
> be building sufficient user needs to drive development of tools that
> would make designing and capturing structured information approachable
> for a wide range of users, that critical mass is going to be diverted
> by the apparent simplicity of using tools that produce Acrobat.

It all depends on what someone is trying to do, and the costs and benefits

Amanda Walker
InterCon Systems Corporation