[
Lists Home |
Date Index |
Thread Index
]
At 01:19 PM 7/8/2002 -0400, Simon St.Laurent wrote:
>At 11:24 AM 7/8/2002 -0400, Jonathan Robie wrote:
>
>>I agree that XQuery is beyond the ability of most programmers to
>>implement by themselves. One estimate I heard recently was that it takes
>>6 *very* good people about 6 months. But how long would it take most
>>people to write compilers or interpreters for other programming languages
>>or query languages?
>
>Most programming languages and query languages have multiple
>implementations.
Of course. Here is a list of 15 implementations and 3 partial implementations:
http://www.w3.org/XML/Query.html#products
These are at various levels of completeness, of course, but for something
that is still just a Working Draft, I think that's pretty good:
>Few have as many implementations as XML 1.0, and there are lots of reasons
>for that.
Of course - but is this relevant?
>>I don't think that a query processor is as simple to write as an XML
>>parser. That doesn't mean that queries are bad.
>
>You seem utterly unwilling to acknowledge that the style of processing
>you're pushing with XQuery is very very different from the style of
>processing that takes place in parsing XML documents.
Huh?
Parsing and querying are two completely different things. They aren't
supposed to be the same. They aren't supposed to use the same technology.
That's why query optimization is not based on BNFs.
>I'm not sure this is worth arguing any more, but I have to say that I have
>less and less sympathy for W3C XML Query (and W3C XML Schema) every time
>this wheel spins around. At this point, I have no confidence that the W3C
>or its working groups have any understanding of why XML was attractive in
>the first place or what it's actually good for. It seems to have gotten
>lost in the roar of vendors rushing to pile on tools for programmers who
>can't adjust to a slightly different view of information.
There really might be more than one opinion of what XML is good for - and
in fact, XML is good for more than one thing.
Regardless, the best way to get a feel for what makes XQuery useful and
attractive may be to download one of the more complete implementations and
play with it. It takes time to grasp a new technology. And then perhaps the
next time around we can discuss XQuery in terms of the language itself,
with sample queries and concrete scenarios.
Jonathan
|
