XML standards coherency and so forth

[Rob Schoening]

> Excuse me if they're naive.

Not at all.

> Firstly, the idea of xml needing a "killer app". I thought xml _is_ the
> killer app?

To the minds of the people reading this list, it probably is.  But the
subscription to this list isn't a good cross section of *any* population!
;-)

XML is certainly the enabling technology for a lot of potential "killer"
information systems.  Unfortunately, it is applicable to so many different
problem domains that it lacks the punch of a killer app.  The story of
Visicalc is that when shown to the computer geeks, they said "computers can
already do that" and when shown to accountants, they said "this is it!"  Now
it seems that the roles are reversed.  The geeks are saying about XML, "this
is it!" and the rest of the world is saying "computers can already do that."
I believe that Java would have followed the same path were it not for the
gratuitous applets that served no practical purpose whatsoever.  Without the
applet, the world would have said "computers can already do that...and 10
times faster to boot".  But once you play tic tac toe on a web page, it is
patently obvious that "computers *couldn't* already do that."


>It sure seems to be being adopted as fast of or in advance of
> released standards for a lot of different products (Ms Word's
> adoptation of
> xml in it's next release comes to mind as a non-web product).

This is, IMHO, a really big deal.  I've seen various figures that report 80%
of corporate data resides on the filesystem of Windows desktops in the form
of Excel and Word documents.  The number of people that spend 40 hour weeks
creating such documents only to have them printed and filed away in the
departmental closet is staggering.  What is more staggering is that this is
fantastic information that is completely inaccessible to almost anyone
except those who generated the data.  It doesn't surprise me one bit that MS
is pursuing XML as hard as they are.  If that data were less opaque, the
value of the MS products would rise significantly.  If cubicle farms could
be mined for data, that would place XML in the "this is it" category.

> Third, about patterns etc. Um... isn't a dtd a pattern for a
> collection/class of documents that adhere to that dtd if they are valid
> instances of it?

Yes, but there is no convenient way to apply logic to these collections.
Set-building among collections of XML documents is very difficult.
Set-building is the mainstay of corporate information systems: reports
reports reports.  XQL isn't much help either.  It's great for extracting
information from a few large documents.  But it doesn't do much good for
sifting through enormous numbers of small documents.  That's great for
SGML-ish applications, but for lightweight documents that represent
transactions, the utility is wasted.

> I recall when rdms' where the ugly step children due to the overhead their
> use entailed and a lack of understanding of their benefits etc. Now they
> are the entrenched majority trying to fend off the next inovation and
> maintain the status quo. Deja vu all over again.

Is there a clear "threat" to the RDBMs?  OO databases are maturing, but they
suffer from the same problem as I mentioned previously: reporting!  Set
processing is very difficult since the stored objects are opaque. OODBMS are
great for the OO programmer, but not so for the CIO.  I can't imagine any
system that can process sets of documents that wasn't (at the lowest levels)
a set-theoretic RDBMS.  (Please, prove me wrong!!) I'm really interested to
see Oracle's forthcoming XML support.  If XML was a neo-first-class datatype
for a RDBMS, I can see fantastic things to come. But if the XML data is
stored as a blob and run through a parser to and fro, it will be a complete
bust.  I'm not holding my breath just yet.

>
> Now, permit me to digress a bit. Having been studying and working with xml
> for the last 18 months or so, it seems to me this market is
> coming together
> very slowly. There are not yet any decent (at least not inexpensive) tools
> out for the  average consumer-programmer that seem worth a darn. I
> implemented a small xml application for a commercial product, and we where
> unable to find usable off the shelf tools at any (reasonable) cost. Ditto
> for commercial C++ drop in parsing tools.

I agree and I also understand that this is the responsibility of the people
and organizations represented on this list.  Or perhaps we need an
xml-app-dev list???  Or some kind of grass-roots organization that can
affect some kind of change....


Rob Schoening


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)