XQuery as a general data processing language WAS: [xquery-talk] XQuery and Web 2.0

Peter Coppens pc.subscriptions at gmail.com
Sat Apr 26 01:48:29 PDT 2008

Sure...no disagreement here....perhaps that just means that a "general  
data processing language" as is the subject of this thread is not  
there yet?

If mainstream adoption of XQuery is what we all care about in this  
discussion (and I might be confusing web 2 mashups, as originally  
intended, with mainstream here), I still think that better integration  
in your more typical application environment (which for better or  
worse looks like it will be oo based for the coming years) is going to  
be key.

Applications like Markmail or whatever other killer app one comes up  
with is a necessary trigger to get the attention of the average  
buzzword allergic and essentially lazy developer, but I doubt that  
you'll keep their attention once they find out what things they have  
to go through to take an XQuery result and integrate it in their  
familiar environment (or vice versa).

What do you think about e.g. LINQ which from what I understand of it  
seem to try and deal with this integration/mismatch problem? Would  
that not better fit better under the "general data processing"  
language umbrella?


On 26 Apr 2008, at 00:23, Jason Hunter wrote:

> Peter Coppens wrote:
>> If XQuery has the ambition to become a general (data) processing  
>> language it will have to integrate (or work) seamlessly with <your  
>> preferred oo programming language> object model.
> I see XQuery as a general _content_ processing language.  Emails are  
> content.  So are books, magazines, tutorials, lesson plans, flight  
> manuals, governmental records, IRS filings, and so on.
> These things fit poorly into a classical object model.
> Question: How do you model a mixed content article in Java?
> Answer: Not very effectively.  :)
> Most people think of content as something you hold as a blob and  
> access via its metadata or simple textual search.  If instead you  
> think of content as a rich description language, then you will see  
> the world like we do on MarkMail.  It's about content, not data.   
> And putting Java in the picture would be a royal pain because mixed  
> content things like email aren't nearly as effectively modeled in  
> Java as they are in XML.
> This is why Daniela is onto something.  Most Web 2.0 apps are  
> content apps, not data apps.

> -jh-

More information about the talk mailing list