Re: No IMG in FIG

Ping writes:
>On Wed, 2 Aug 1995, Michael Johnson replied:
>> No, the content model of <P> is too restrictive to be used as the content
>> model of <FIG>. The respective content models are more useful as they are,
>> and I personally don't see any problem with simplicity or consistency either.
>
>Please point out how you mean that the content model of <P> is "too
>restrictive" to be used for <FIG>.  I am asking primarily that we
>remove the restriction that prevents <IMG> (and <FIG>) tags from being
>used within a <FIG>.  Every added restriction makes HTML less simple.

I agree that <IMG> should be allowed inside <FIG> but that is not what I was
objecting to. Consider the fact that <P> cannot contain lists, <PRE>, <FN>,
<NOTE> or <P>. <FIG> needs to be able to contain (almost?) anything that can
be contained in <BODY> or it is less useful than it could be.

<P> on the other hand should not be containing things which are not
semantically part of a paragraph.

>> Don't take this personally, Ping, but it seems to me that you have been
>> making a large number of poorly thought out proposals lately. You seem to
>> have a less than perfect understanding of the mission of HTML or else you
>> just aren't taking the time to really consider the consequences of some of
>> what you propose.
>
>I can certainly accept this, but a general opinion on the quality of my
>suggestions doesn't help us to work towards the goal of a better markup
>language.  I'd appreciate much more an explicit description of how my
>understanding is flawed, specifically what parts of which proposals you
>feel to be poorly thought-out, and what ill consequences may arise.

Well, for an example, you proposed making the <FIG> content model match that
of <P> which if you had done a little research and thinking on, you would
have realized was not such a hot idea, as I explained above.

As another example, you proposed a while back that <FIG> should be allowed
inside of <P> which turned out to have ramifications that you didn't
consider, mainly again a matter of conflicting content models.

It sounds like you are forming an idea in your head that all HTML containers
should have the same content model (I'm inferring your thought processes from
what you have written) based on the idea that this would be simpler and that
would make it better.

This would be a bad idea. It's hard for me to express why this is, but
basically it would destroy the semantic meaning of the various HTML container
elements. A paragraph is not the same thing as a list item which is not the
same thing as a heading, which is not the same thing as preformatted text, and
the content models of these things should reflect their nature.

I don't know what kind of background you have in text markup but my impression
is that you don't truly understand the philosophy behind generalized markup
languages. It appears to me that you are used to tools that compose documents
according to their appearance rather than their content. It appears to me that
you are unconciously still applying a presentation-oriented mental model to
your ideas for HTML.

I may of course be way off base, but since you asked, I thought I would try
to provide a constructive answer.

One of the big problems facing the HTML community right now is the damage that
has been done by well-meaning people who tried to treat HTML as a presentation
language (or turn it into one) rather than treat it as what it is, a language
for describing the structure and content of a document.

If I am wrong then I humbly apologize.

Michael Johnson
Relay Technology, Inc.

Received on Thursday, 3 August 1995 07:48:22 UTC