Re: [ba-ohs-talk] Using Java in XSLT
On Saturday, December 29, 2001, at 09:24 AM, G. Ken Holman wrote:
> I think the breakdown will be split differently.
>
> I think the use of XSLT for "pull-oriented" transformations will not
> only diminish, but probably disappear with XQuery, but that XQuery will
> not accommodate the "push-oriented" transformations that are sometimes
> required.
>
> Granted, document people more often need push-oriented solutions than
> data people, and data people can use pull-oriented for most of their
> work ... so the end result is as you say, but not for the nature of the
> data, but the nature of how the data is processed. (01)
The nature of the data usually determines the way that the data is
processed. This is a fact that the OHS system should embrace with open
arms. So what I said stands, but as a simplification of the overall
state of things. (02)
If people like Evan Lenz have their way, (those who are completely
pissed off with the way XQuery has developed), XSLT will remain the
de-facto standard for pull based querying, and eliminate XQuery
completely. If people like Jonathan Robie have their way, XQuery will
completely eclipse XSLT, which will make a prettier bullet point on
one's resume. Neither will win out in the end, of course, because both
address different problems with substantial, but necessary overlap into
each others turf. (03)
Both languages have their strengths and weaknesses, and their own
rallying camps, but there is consolidation happening which should lessen
the bloodshed. Specifically, XPath specification is shared between the
two technologies, so the foundation for node selection is common. (04)
On the other hand, XQuery's scope is fairly broad, and will easily
support both push and pull metaphors. Some of the products and services
coming down the pipe that I've been privy to in my recent job
interviewing will demonstrate this quite nicely. (05)
Then of course, there are people like me... The ones who think that 90%
of the stuff that the W3C produces these days is politically motivated,
and bureaucratically determined garbage, that should have come from the
real world, rather than the abrasive collaboration of academia and
big-business. But that's just me. (06)
--
Tom Bradford - http://www.tbradford.org
Chief Architect - The dbXML Group, L.L.C.
Developer - Apache Xindice (formerly dbXML)
Maintainer - jEdit-Syntax Java Editing Bean
Co-Author - O'Reilly's "Learning Xindice" (07)