Re: What are the problems with IDML? (fwd)

Murray Altheim (
Thu, 22 Aug 1996 11:38:53 -0500

Message-Id: <v02110104ae3eb0af63ac@[]>
Date: Thu, 22 Aug 1996 11:38:53 -0500
To: MegaZone <>
From: (Murray Altheim)
Subject: Re: What are the problems with IDML? (fwd)
Cc: Doug Donohoe <>,

MegaZone <> writes:
>>> 2.META tags are poorly suited to specifying products. Some
>>>     Identify merchants have over 600,000 products; to catalog their
>>>     products using only META is ugly and impractical. We found
>>>     that a separate, dedicated tag just for product-tagging and
>>>     content-tagging provided greater flexibility and clarity.
>>> I don't get this, I really don't.  I've looked at their tags and I
>>> don't see anything I can't do with META.
>>This is merely an elaboration of the previous point.  You say "I
>>don't get this".  What don't you understand?  Perhaps we can clear
>I don't see any reason to use your tags when META does the job just fine,
>is valid, and is proven.
>>All I ask is this:  Sit down and think what it would really take
>>to represent 100 products in a manner that a robot could understand.
>I have, I've done things like that.  All it would take is a good naming
>convention.  I must be missing something because I sure can't see anything
>provided by IDML that a well constructed naming convention wouldn't provide
>in META.

Well, I'm mystified. Either this is a silly argument or I'm off my nut. I
think Dan answered this but I'm still seeing later notes continuing into
the ozone.

To repeat Dan's message: META is only allowed in HEAD, regardless of
whether or not the HEAD _tags_ exist. As proooof:

   1. TITLE is required.
   2. TITLE is _always_ within HEAD.
   3. Therefore, HEAD is always present by implication.

   4. The first instance of non-HEAD content begins BODY,
      regardless of whether a <BODY> tag exists or not.
   5. Therefore, BODY is always present by implication.

   6. META only exists in HEAD.
   7. Therefore, document content (ie., within BODY)
      cannot be marked up with META.

[My unassailable logic baffles even me. I will next prove life exists on Mars.]

As Doug points out, if a single document describes multiple products, then
providing information about the individual products by adding that
information to the HEAD of the document is not only "ugly and impractical",
it simply won't work in a real system, unless one resorts to separating
each product's "meta-information" to the HEAD of the document and providing
some link between the META content and the actual document content. This
could be done by lobbying for IDREF/IDREFS in META, but to what end?

If IDML allows document content (within BODY) to be marked up with product
identification information (not _document_ metainformation), then it seems
that IDML is certainly a viable markup approach. This regardless of its
other merits/demerits.

I don't think industry-specific HTML extensions would probably be adopted
as standards or supported in products outside of the target industry, much
less standardized by IETF, ISO, etc. You might look within your own
industry to propose its own standards.

Aside from standards, you just need a supporting DTD to be "SGML legal."
This could be done with a modular DTD, with IDML as a module, similar to
the chemical industry's molecular modelling [CML] DTD that incorporates
HTML. You'd simply create a DTD with an FPI something akin to:

     "-//Identify//DTD HTML 3.2 IDML Extensions//EN"

I don't see Netscape, Microsoft, Spyglass, etc. providing HTML browser
customization for each industry -- too much specialization and too many
industries. I see this only happening with an "open SGML" Web (where you
simply provide the DTD for your specialized document type).

But why would you need browser customization? If IDML is truly intended to
be transparent to browsers, and only important to search engines, you might
contact several of the search engine companies and find out how they might
be able to work with you on a solution that works within the rest of the
Web. It could be a selling point for them. You've obviously done that on
the Identify site, but wider utility would not require the custom Java
agent or any specialized searcher; users could use their current search


[CML] Chemical MIME and Chemical Markup Language (CML)
      Peter Murray-Rust, <>

     Murray Altheim, Program Manager
     Spyglass, Inc., Cambridge, Massachusetts
     email: <>
     http:  <>
            "Give a monkey the tools and he'll eventually build a typewriter."