On Wed, Jun 06, 2007 at 11:58:30AM -0000, De Zeurkous wrote:
> Like the IBM PC, those OSes are either obsolescent or outright obsolete.

I think you're talking to the wrong crowd about "obsolescent" or
"outright obsolete".  :-)

> > other BSD installs,
> This may sound incredibly naive, but we could just agree upon one
> disklabel format, right?!

If you're not familiar with the term "bikeshed", I refer you to the
search engine of your choice.

> I'm not proposing we dump wedges -- I'm simply stating that it is a
> solution to compatibility problems between BSD, the host architecture,
> and/or other OSes. It's not a nice replacement for our current layout for
> physical devices.

You state that "it's not [...] nice", but I don't think you've really
stated why.

> > And then there's the question of
> > how many partitions does that disklabel/port support?  8?  16?
> 
> Isn't this even a consistent number? :S Of this I was not aware. Another
> reason for a new, unified disklabel.

And a reason for you to do a bit more research to better justify your
statements / positions.

> > It's a "flag day" for a port if/when it changes [...]
> Then again, this should have been fixed a long, long time ago; preferably
> that mistake should never have been made...

True.  But not particularly helpful except to help us learn from
that history.

> > But GPT is an existing format that maps well to wedges,
> > which we want for other reasons.
> Is some brain-dead existing format better than a new format that is
> guaranteed to work?

Why is a new format "guaranteed to work"?  My experience has been
that it's actually difficult to see into the future and get everything
right.  And why is GPT "brain-dead"? I recently implemented a pretty
basic GPT setup for hire, and it didn't seem all that brain-dead
to me.  It seems well-designed for its needs, and not over-engineered
(like many things in this industry can be!).

> If so, we should consider implementing a full-fledged
> ROM BASIC-clone in the kernel debugger...

Now that sounds like fun...  Thanks for the suggestion!
( not that I would expect it to be accepted into the tree, but it might
  be a nice diversion...  :-)

> > and more of an issue for USB or
> > firewire (or eSATA) disks that may well be shared between different
> > kinds of systems for backups or sharing data or whatever.
> 
> We have something called a 'file server' for USB and SATA.

So you're suggesting that we put USB sticks on a Windows (or Mac or
Linux or FreeBSD or other) system and share the device for us to use
them on NetBSD?  That seems less than useful.
I mean that if I have a USB stick formatted for use on one system,
I should be able to walk over to any other system with USB (be it
Mac, Windows, Linux, DragonflyBSD, MorphOS, or whatever) and have
it "just work".  Perhaps with partitions for different filesystem
types.  Perhaps with one common filesystem.

In any case, I do believe this is a discussion that we've had in
the past.  Please read the archives.  If you still want to discuss
it after reading the archives, then please enumerate the technical
issues you have with the direction that the project has chosen to
take.

-allen

-- 
Allen Briggs  |  http://www.ninthwonder.com/~briggs/  |  briggs@ninthwonder.com