[xquery-talk] XQuery spec and timeframes
fcohen at rainingdata.com
Mon Jan 23 15:54:12 PST 2006
Hi Ken: When I read your message I hear pride, anger, fear, and
hesitation. I've been there too. It's not an easy thing to build a
new standard and I admire what you and others have done to get XQuery
this far. It's a good accomplishment.
Getting the XQuery spec to a 1.0 final status is very important to
two groups: developers and managers.
Last year at the SDForum.org Web Services Conference Tim Bray and
Adam Bosworth talked about XQuery as a cake that took too long to
bake. The audience was Java developers. Trying to fight against that
attitude from the thought leaders is hopeless while the spec is not
Secondly, the enterprise and institutional software architects and
program managers I am calling on like the promise of XQuery and are
ready to rally around the new standard. But they wont accept "the
spec is basically final" or "the spec does not include xxx and we
have added it to our implementation." (The xxx could be full text
search, try/catch/exceptions, or any number of other enhancements.)
They need the 1.0 spec to defend their choice of a new platform.
XQuery 1.0 with kinks still inside is fine with me rather than
waiting any longer.
About XML-QL and XQL adoption... XQuery is a native XML programming
language and Java is an object-oriented programming language. They
are natural competitors. I have no reason to think that the object
bigots in the Java community are motivated to support XQuery as
anything more than an improved SQL language. The win for XQuery is
with developers that are working with complex XML data. To them
XQuery is a winner for system performance and scalability and rapid
development of XML services.
On Jan 23, 2006, at 2:27 PM, Ken North wrote:
>> With XQuery we're not talking about a protocol for document
>> schemas, we're talking about a spec for a language.
> We have a different view of whether SGML is a language, so consider
> examples. The timeline for Ada was 20 years (1975-1995). The
> timeline for SQL
> was 13 years.
> Unlike hardware technology, there's no Moore's Law to ensure
> developing a
> standard today will be much faster than 10, 20 or 40 years ago.
> Human behavior
> hasn't changed that much. People today still have similar demands
> on their time
> (other duties, families, etc.).
> Perhaps the real topic for discussion is whether finalizing a spec
> quickly is of
> greater value than ironing out the kinks and developing consensus.
> Look at the XML-QL and XQL query languages. Java implementations of
> XQL were
> available quickly (compared to XQuery), but look at adoption rates.
> building XQL (or XML-QL) into their plans today?
> There's another example that's closer to home. Raining Data's database
> technology, originally developed by Dick Pick and Don Nelson, was
> available long
> before Microsoft and Oracle even existed. But those companies bet a
> (SQL) would drive the database market. Dick Pick's companies
> (Microdata, Pick
> Systems) took a proprietary route with its DBMS, OS and query
> The standard paid off for IBM, Microsoft, Oracle and others. So
> this time around
> it's a good idea Raining Data is building to standards, even though
> the XQuery
> timeline may be disappointing.
>> If Java took this long we would all be programming in C today - or
> Not a good comparison. The team that developed the Oak language
> spec worked for
> Sun's Green project. They were locked away for 18 months and had no
> The 9-person committee that created COBOL included people from
> organizations. It took only six months (April 1959-December 1959)
> but the people
> were working almost full time on the spec.
> talk at xquery.com
More information about the talk