XQuery as a general data processing language WAS: XQuery and Web 2.0

[Jason Hunter]

Thomas Lord wrote:
> Jason Hunter wrote:
>> http://xqzone.marklogic.com/pubs/3.2/apidocs/All.html
>>
>> FWIW, it seems messy and unnecessary to me to change the language in 
>> any significant way to support these things.  I prefer the Java 
>> approach. Simple language, complex libraries.  Empirically you can, 
>> for example, do fast text search through library calls.  So let's do 
>> that.
> 
> Do you guys augment XQuery with a lot of non-standard functions?   I can 
> see why people do (even if you don't) but it's a trend I'd like to try 
> to resist and suggest others resist, for obvious reasons.

Yes, XQuery 1.0 leaves many critical pieces out of the standard.  Thus 
we augment the language with functions.  Our goal with MarkLogic is to 
be useful and salable, not academic.

My point from earlier is I'd rather the critical pieces became standard 
via function libraries (like servlets were for Java) than as core 
language changes.

> One alternative is to add just one capability: the ability to call out 
> to other (possibly "closely connected") net services in a uniform way.   
> That way you can substitute implementations of the extra functionality 
> independently of substituting implementations of XQuery.

I like calling out to external services.  I created MLJAM to do that on 
MarkLogic:

http://developer.marklogic.com/howto/tutorials/2006-05-mljam.xqy

But of course when the service in question is fast indexed text search, 
it's not the kind of thing you can do as a net service.  That needs to 
be in the product's core.

-jh-