[
Lists Home |
Date Index |
Thread Index
]
Peter.Hunsberger@stjude.org (Hunsberger, Peter) writes:
>Sure, if all you know is that you use
>
>document/foo[3]/bar[@blah='gunk']/@price
>
>to get the value you shove into column 12 of table 87 then good luck.
>My point is that if you expect any of this to work automatically then
>you're going to have to be willing to do the work of getting a real
>schemas. Today you have the XSLT++ produced by Frank in IT that spits
>out your spread sheet. Tomorrow, maybe, you'll need the xsd++ that you
>feed into the magic application that then spits out your spread
>sheet...
Perhaps. My hope long term is that such schema development will be
something done by more ordinary "knowledge workers", not "Frank in IT",
and that it will be done iteratively, as workers process messages and
their mappings and meanings in such a way that the computer can deal
with similar circumstances in the future.
Not AI (or even "automatic") by any means, but a long way from setting
the rules in advance by committee and setting up processing around those
expectations exclusively. XML makes continued human involvement in such
things quite plausible from my perspective.
--
Simon St.Laurent
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!
http://simonstl.com -- http://monasticxml.org
|
