[
Lists Home |
Date Index |
Thread Index
]
On Sun, 12 Dec 2004 12:04:00 -0500, Tom Bradford
<tbradford@openlinksw.com> wrote:
> My guess is that the reason why XQuery hasn't yet taken the world by
> storm is that it hasn't overwhelmingly made its case as to why it's
> better than XSLT.
Has XSLT taken the world by storm? My sense is that it is a bit like
Perl -- there is a community of ubergeeks who know it well, and the
ordinary mortals depend on them for help when called upon to do
something at all challenging. One early hope for XQuery was that it
could handle most of the work that people did with XSLT *and* ordinary
mortals who understand SQL, scripting languages, etc. could become
facile with it. I see very little sign of that actually happening.
> Personally, I like the declarative model, but the vast majority of
> developers don't, which leaves me bewildered as to why XSLT still
> remains more popular than XQuery. Maybe it's just a matter of properly
> marketing the two solutions to their target audiences instead of making
> them available as a mish mosh of bloatage, and hoping that the consumer
> will be able to sort it all out for themselves?
The biggest reason is probably that XQuery is still not a
Recommendation. Major vendors are going to be very cautious about
fully supporting it until they know that the spec is cast in stone.
The memory of XSLT 1.0 is a bitter one ... (breaking changes were
introduced very late in the process).
XSLT may well have missed its window of opportunity to move beyond its
core niche as a query language for XML repositories. XSLT 1.0
extensions to make it more usable and the tools that support it have
matured, and the APIs and programming languages with built-in support
for XML have evolved rapidly in the 5 years that XQuery has been in
progress.
Finally it's just not clear to me that XQuery really does what XSLT
does in a developer-friendly way. It's certainly a lot more
approachable as a way to query databases than XSLT, but is it
dramatically more convenient as a way to filter, merge, and transform
XML data and services? Is anyone even marketing it that way any more?
Most of the commercial interest in XQuery these days is as a database
interface,
In the long run of course the consumers will sort it all out for
themselves and all we are saying here will be forgotten. Still, I
wonder what people think will eventually fill the niche that XQuery
was targeted at (something like "the power of XSLT and the
developer-friendliness of SQL and script"). Some possibilities:
1 - XQuery (or maybe some profile) will arise from the bloatage and
fufill its destiny
2 - XSLT 2's power will motivate mainstream developers to learn to grok it
3 - XPath-aware procedural XML APIs will mature to handle the job more easily
4 - XML data will become a first-class citizen of mainstream
programming languages.
I'll predict 4) as where it will all settle out in 10 years; 2-3 are
certainly happening to some extent and will be important in the
interim. I'm keeping an open mind about 1 - XQuery certainly has a
role to play as a query language (rumor has it that SQLX will
incorporate it by reference?) but I have no idea whether it will move
out of that niche.
|
