Using Python for processing of large datasets (convincing managment)

[Anders Dahlberg]

"Thomas Jensen" <spam at ob_scure.dk> wrote in message
news:3D282290.1050101 at ob_scure.dk...
> Anders Dahlberg wrote:
>
> [snip]
>
> > Considering it's compiling down to java-bytecode which are then run with
a
> > java-vm - It should scale as well as any java solution (or atleast
~90%).
> > Links could be:
> >
> > http://java.sun.com ;)
> > http://www.bea.com/products/weblogic/jrockit/index.shtml - very good
server
> > vm. (optimised for intel)
>
> [snip]
>
> > disclaimer: This is just what I believe - if you want to base your case
on
> > this I believe you should contact other companies using java with
> > multiple-cpus!
>
> I am quite certain that scaling (well) to multiple CPU requires one to
> use threading at least.

Yes, this should be quite straigthforward to implement.

> Scaling to several physical machines might be
> (relativly) easy with Java, but imagine it must require some coding, or?

Yes, java is no magic bullet :(

I *imagine* it would be "easy" due to java-rmi protocoll and maybe even a
technology as javaspaces (distrubted space - "linda"?).
E.G. http://www.javaworld.com/javaworld/jw-10-2000/jw-1002-jiniology-p5.html
(it's not python - but should be similar to a jython implementation)

> --
> Best Regards
> Thomas Jensen
> (remove underscore in email address to mail me)
/Anders