[
Lists Home |
Date Index |
Thread Index
]
I have nothing to disagree with in your response. I often work from the
practical result back to the theory, to derive useful theorems that I can
later apply to other problems. But my experience is that this is not the
best way to develop applications for production use in the wild as several
round trips may be necessary (as has been seen, and will continue to be
seen, I think, in the XML community).
I will call, again, for specific proofs people already have, or for
requests for very specific proofs, that people need right now, or expect to
need in the near term.
Perhaps if we take this discussion up to a higher level of academic
achievement, it can progress and be fruitful.
Personally, I would like to see XML proofs in line with Knuth's work and
Codd's. Proofs that justify its existence, and define its scope and
usefulness.
Something other than what I often hear... "but the XML document can be
__anything__ you want it to be !!!!!!" as if that was proof enuf. When I
respond that a Word doc (using Microsoft's word document markup language
and symbol interpreter) can be anything you want it to be, the XML advocate
usually just looks at me in total amazement.
Thank you.
At 12:32 PM 8/25/2003 -0500, Bullard, Claude L (Len) wrote:
>Yes, once one has a good idea of what needs proving.
>Otherwise, like proving that a bumblebee can't fly,
>one is modeling from axioms, and the bee just goes
>about its business doing what it will. Abduct, induct,
>then deduct.
>
>Markup scales where it needs to first, in the applications
>which its inventors had in mind. Note, DTDs were added
>later. Formalization was done later. This is historical
>even if not necessary. In short, don't wait on proofs
>nor ignore them when available. Just be sure of what is
>being proved and that it is relevant to the task.
>
>To the point: is a wall-to-wall XML database the right
>solution for any problem? No. Can we define in advance
>all of the problems it is or is not applicable to? Not
>all. I agree with you 100% on that.
>
>As noted elsewhere, when looking for the proofs, it is
>not XML that needs to be compared, but the data models.
>As others have noted, there is lots of work being done
>on these. Meanwhile, hybrids rule the niche today.
>
>I do note that the environment for which XML, (not markup)
>was designed or adopted from SGML, is one of decentralized
>and loosely coupled sources, not one where normalization
>is the norm. Will that create update problems? You bet.
>Like 404, it is a cost of using the system. The only
>proof 404 needed was 50 years of trying to get around
>it. Progress in fielding very large distributed hypermedia
>systems was made only when that constraint was relaxed.
>The way around the update problems so far is hybridization
>and tightening the coupling. The ideal of full decoupling
>is not just risky, it is unworkable to date.
>
>You're right about Curie. One shouldn't bet the farm
>for a prize in husbandry, or die of the experiment,
>but there would be no Wright Brothers without Lilienthal.
>
>len
>
>
>From: lbradshaw@dbex.com [mailto:lbradshaw@dbex.com]
>
>Awww, now, gee whiz..... doing a math proof can be as exciting and
>exhilarating as any other form of discovery, and doing so first can save
>one many stubbed toes, later.
>
>After all, Codd's proofs led the way, as did Knuth's, and were not derived
>from existing advanced art but from theory and science. While it probably
>goes both ways, some proofs coming from experience and others derived
>purely from theory. saying that waiting on proofs makes me a cave man who
>is frightened of tomorrow is just personal.
>
>Which is something I will not respond to, :), other than to say that if
>Curie had done the math, used the full scientific method, waited for
>results and included advancements from other scientists, maybe she would
>not have died of radiation poisoning.
>
>Thanks!
>
>-----------------------------------------------------------------
>The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
>initiative of OASIS <http://www.oasis-open.org>
>
>The list archives are at http://lists.xml.org/archives/xml-dev/
>
>To subscribe or unsubscribe from this list use the subscription
>manager: <http://lists.xml.org/ob/adm.pl>
|
