On May 24, 6:08 pm, Volker Hetzer <firstname.lastn..._at_ieee.org> wrote:
> DA Morgan schrieb:> I see no reason why vendors should agree on a procedural language. If
> > that were a requirement IBM and Microsoft could hold Oracle hostage to
> > their lesser capabilities. No FORALL, no user defined operators, no user
> > defined type bodies, etc. That is not in the interest of the customers
> > who benefit from the competitive environment where you guys get to play
> > catch-up and Oracle is forced to aggressively enhance their products to
> > maintain their lead.
>
> > Viva la difference.
>
> Great. VHS versus betamax, HD DVD versus blue ray and so on.
> That's exactly the reason why logic moves out of the database.
> People need applications but aren't willing to have a bunch of different,
> loosely (if at all) connected databases, depending on what the application
> vendor happens to like.
> And instead of of agreeing on a standard, the database vendors continue
> to force application developers to waste time supporting wildly different
> languages. End result -> Java stored procedures. Great.
> No whatever features? Somehow compiler vendors managed to get C++ supported,
> right? It's not perfect but its way easier to port from one C++ compiler to
> another than from one database procedural language to another.
> Ditto for sql, but for pl/sql, standardization is naturally wrong, just as it
> is naturally right for just about everything else.
>
> Right now, XML is the coming thing, with a standardized query language. Sure
> it's committee designed and you can lift your nose at it but there's a
> database vendor independent query language, XQuery and everyone goes
> and uses it on every database that supports it.
> After that, who needs data warehousing when search engines just need a bit of
> additional processing capabilities to make the next inroad into traditional
> database territory?
>
> Because proprietary database programming languages just don't do?
>
> Sorry, but "I see no reason why vendors should agree on a procedural
> language." is really low.
>
> Volker
> --
> For email replies, please substitute the obvious.
You think XQuery is the next thing. Try Semantic Web SPARQL.
XQuery... Ha! :)
Received on Fri May 25 2007 - 09:27:34 CDT
