Re: No future for DB2

"Serge Rielau" <srielau_at_ca.ibm.com> wrote in message news:3kug55F10ekpvU1_at_individual.net...
> Noons wrote:
> > Sure. But there is still a very clear limit: going back to
> > the volumes bit I mentioned above. Those of us who still have
> > to process serious amounts of data in as little time as possible
> > are still using good old C/C++, shell scripts, perl here and there
> > and maybe in a moment of light-headed enthusiasm, python/php. (Now,
> > THOSE are interesting languages!)
> Funny, I thought assembly language to rules supreme.
> Layers of abstraction:
> microcoding,
> assembly,
> C/C++
> Java
> ...
>
> Folks are debating DBA skills in these forums. Ever tried to get C/C++
> skills? It's a b**** to get a good C programmer nowadays.
> Certainty the kids who wrote C when they were 16 and hit the job with 10
> years experience don't exist anymore.

Well, I fit that description -- I guess that's why IBM hired me!

> Point being, you wouldn't go back to micro-coding, you wouldn't go back
> to assembly. Too expensive in labour, too unwieldy.
> C/C++ is already turning into a niche skill.
> The abstraction layer just keeps on rising because Moore's law and
> inflation favours throwing Hz and bytes at any problem over brain.

It's also a management issue. Managing 10Mloc of C code is tedious. But a set of 1000 Java "objects" is somewhat easier, at least from a conceptual level. This is the force that is driving Java app servers with JSP/servlet technology in the industry today, and is the reason why WebSphere and the Apache group's products (Jakarta, Tomcat, etc) and countable others exist. XML fits in here in a simliar vein but is being driven into the DBMS products directly instead of being handled by a separate product that separates the presentation from the data.

--
Matt Emmerton