Jason Thorpe <thorpej@nas.nasa.gov> writes:
>On Thu, 3 Apr 1997 22:54:18 +0200 (MET DST) 
> Lennart Augustsson <augustss@cs.chalmers.se> wrote:
> > In general, wouldn't it be a good idea to reserve the device numbers
> > in -current for the most common add-on devices?  This would make it
> > simpler for those of us that use several.

>...for atapi, probably, since the interface that code uses won't change
>when it's integrated...

It won't?  I seem to remember that the last time the integration of the
ATAPI code came up on the public mailing lists, it was argued that the
code should be changed to provide a SCSI-controler like interface so
that it could use the standard SCSI subsystem instead of duplicating large
parts of that subsystem's code.

>the pcmcia code, however, might, so I'm hesitant
>to assign a major for it.

Since you only have to mark it reserved in a text file in "doc", I
don't see just what the problem with that is.

If you're worried that you might allocate major numbers that will never
get used, just put an "expires `some future date'" on the relevant line
and make the assigned party provide some proof of progress when that
date is reached.

If this gets implemented, it would be nice if a reasonable number
of major numbers could be set asside for "local use".

-- 
Christoph Badura

Now available in print: Lion's Commentary on UNIX 6th Edition, with Source Code
			http://www.peer-to-peer.com/