> From firstname.lastname@example.org Thu Oct 2 12:41:23 1997
> Content-Return: allowed
> Subject: Like your XML article
> To: "'email@example.com'" <firstname.lastname@example.org>
> X-Mailer: Internet Mail Service (5.0.1458.49)
> I just got done reading your "The Evolution of Web Documents: The Ascent
> of XML". It begs the question however, "Where is the Web in 2 to 3
> years." At our Laboratory we are struggling with whether we invest in
> Web&SGML now or whether SGML will ever be popularized (used widely
> with the web). Do you have an opinion or have links with opinions?
> Gary Danielson
> Pacific Northwest National Laboratory
The bottom line is that, for Web use, XML in 2 or 3 years will be all
the SGML you will ever need. XML discards most of the esoteric and
arcane features of SGML, and keeps the features that Web developers will
Now, there is a question of ramp-up time, as there was with the Web
itself. XML needs to reach that critical mass of acceptance, and I
believe that critical mass is a year or two away: W3C is embracing it as
a Web solution and slowly but surely members of the SGML community are
coming around as well.
So where will the Web be in 2 or 3 years is anyone's guess. But here is
my thought --- that the WWW will evolve into the XXX:
1. The first X is for eXtensible markup. HTML will give way to
specialized XML documents. XML DTDs will provide documents with
verifiable typing when needed, and XML tags themselves will be prime for
automated extraction and processing of data. XML is the right tool for
the job of data annotation and manipulation.
2. The second X is for eXtensible addressing. URLs were great in
binding name to location in a global context. I think that lots of
current efforts in metadata, from the Warwick Convention and the Dublin
Core to the Platform for Internet Content Selction, will give rise to a
good solution (call it XRLs) within or without Universal Resource
Identifiers --- and combining point 1 above with point 2 here will give
rise to a way to bind the "meat data" of 1 with the "meta-data" of 2.
3. The third X is for eXtensible transport. HTTP is going to give
way to its "next generation" progeny, HTTPng. Why not learn the lessons
of PEP and transactional systems, and create XTP, a meme for developing
specialized protocols where needed with respect to state management,
format selection, security, compression, multiplexing, intermediaries
like proxies and filters, "flow types" such as backward-directed or
asynchronous, real-time streaming when needed, content negotiation, and
best of all (and really far-fetched) leveraging off XML and XTP to allow
the automation of tasks previously un-automatable.
I'm sure these thoughts are farfetched --- just the mention of a triad
like XML, XRLs, and XTP would probably get me burned as an iconoclast
(or is it heretic?) --- but you wanted to know what I think. I think
the smart money is on investing in XML. Everything else is up for grabs.
Don't it always seem to go, that you don't know what you've got 'til
-- Janet Jackson sampling Joni Mitchell