>> If you tell me _why_ you think all-in-one-fs is "BAD", what you
>> think is so dreadful about it, I'll be happy to explain why it
>> doesn't always apply.
> Log Files packing the root slice is my #1 answer.

And what's wrong with that?

Seriously, on a single-user machine (as in, one that's used by only a
single person, sysadmin and user rolled into one), it's not at all
implausible that the user/owner/admin wants to be able to trade off
logfiles against other data dynamically, which fixed-size partitions
can't do.

> /var should (in my opinion) *always* be separate from / on a
> production machine.

For what values of "production"?

> Users packing your root slice is #2,

Again, not applicable for a single-user machine - of which there are a
lot these days.

Part of being a good sysadmin is knowing when various filesystem layout
philosophies are appropriate.  Dogmatic insistence on splitting things
up is no better than dogmatic insistence on rolling everything
together.

> /tmp being mounted on another spindle (or, if you can get it, another
> SCSI bus) can make large performance differences.

Yes.  But many people, especially on single-user machines, are willing
to accept the performance lose for the sake of not having to buy
another spindle.

> Some browsers [...] store downloaded files in /tmp until the download
> is complete, and *then* copy them to their destination.  if /tmp is
> mounted in your root slice, and you try to download a large package
> or patch bundle from the web, guess what... you just might pack your
> root slice to 105%.

But if your root filesystem is the same as your every other filesystem,
you won't run out of space until you are absolutely out of space.  None
of this "but I've got half a gig free over here, why can't it use it?";
this is actually a case where one-big-fs *wins*.

>> NetBSD has decided to disagree, and if you disagree with them on
>> this point you may want to go looking for another OS, one better
>> suited to your particular tradeoffs.
> if there were a better one available, I would.  But NetBSD is the top
> of the stack, as far as I'm concerned, and that's why I'm peeved to
> see it going in this direction.

Yeah, I know the feeling.  My own problems with NetBSD are different in
detail, but the same in this regard - it's going in directions I don't
like but is still the best of a bad lot.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B