Beowulf digest, Vol 1 #289 - 11 msgs

[pbn2au]

> I agree with the notion that you eventually have to work (or rather,
> wrestle) with Office products, because the software is in such wide use.
> Office products can be good for prototyping data structure, and generating a
> quick graph if your data is limited in scope and size.  Getting the data to
> and from "Office form" is actually a large task, and full time job for some,
> including yours truly.  I once had to process three gigs of data at once
> using Access, just cause the boss said so, and it wasn't pretty . . . if I
> had access to a cluster at the time, I would have done it much faster.
>

I disagree with the "need" to wrestle with office products. Has everyone forgotten
Star Office? The data can be directly transferred to documents that could be
opened in MS OFFICE if the need arose.

Also a side note I have been getting totally exasperated by the MS clustering
demo package. This thing is a pig, and it builds on quite a bit of the information
discussed here. The white papers are good and if applied to Linux/beowulf
clustering would be acceptable. However, when applied to a cluster of 233's as
part of a dual boot system, it completely bogs down. I'll be reattaching my 486's
and going back to linux thank you.


David Campbell
Campbell Consulting
Middleton Idaho
pbn2au at qwest.net