Benchmarking

[Ivan Voras]

Since I received mostly negative comments on the whole thing, I'm 
considering doing myself a favour and taking it off public access, although 
I *still* think that the benchmark is valid under its goal, and I will 
defend it as such.

But 'm willing to learn :) Just for argument sake, if I ever do something 
like this again, what should I do to make it better? So far, I've got:

- Increase the number of files for bonnie++ (I agree)
- Don't use bonnie++ at all (I disagree - what else to use?)
- Enforce same partition/slice size for NetBSD (I agree)


Until such opportunity, are there any suggestions about what to do what the 
current article?

- Remove bonnie++ filesystem results?
- Remove NetBSD from the article?

What suprises me that nobody's disputing bytebench - I thought its results 
were far more interesting... :)