Mike Cheponis wrote:
> It comes from a deep-seated belief: never impose limitations when
> none are needed.

It comes from deep-seated belief: KISS & keep things under control.

> to do a molecular model that needed 11 GB?  Well, I'd be SOL.

I'm not sure you would be happy with performance of 11 GB
file-based (and possibly fragmented) swap ;-) And I'm cringing
in terror while I image your defragmenter trying to defragment
live swap file ;-)

> I don't understand why everybody seems to want to hardwire constants
> into the filesystems when a more flexible approach doesn't need
> to do that.

The swap file size is not hardwired, if you mean that.

> > On the other hand, if my system was using the
> >"free disk space" for swap instead of a fixed partition, I would
> > run out of swap all the time when my disks become full. ;-)
> 
> This can be easily handled by quotas or other policy decisions.

I'm eager to know how you'd like to enforce quotas on swap ?

> Gentlemen, we simply -must- break out of this 20th-century unix
> box we've
> been living in and re-evaluate how to make our beloved OS the winner in
> the 21st century!

This sentence is a crap.

Jarda