I do think Enon has a point.  It's true, you -can't- determine if what
you're adding to the kernel is safe.  (I suppose you could run the code
in a vmware penalty-box and trap its operation, or maybe use some kind of
code analyzer, but that is going to the extreme.)

Windows reliability has been crappy because the h/w manufacturers
usually did a lousy job of driver writing, and plugging that stuff into
windows often causes windows to be unreliable.  To combat this, Winxp now
uses signed drivers.

Nobody is suggesting that somehow 3rd party drivers are evil and they
are buggy and they are trying to bring down the system or be some kind
of clandestine FBI data-collector, or ....

...but you can't know (unless you're willing to go to extremes, as mentioned
above).


As for announcements, I don't see how they hurt, tho.

-Mike

On Sat, 8 Dec 2001, Aaron J. Grier wrote:

> From: Aaron J. Grier <agrier@poofygoof.com>
> To: netbsd-advocacy@netbsd.org
> Subject: Re: Bad Mistake by Announce List Owner
>
> On Sat, Dec 08, 2001 at 10:23:57PM -0600, Enon Nahca wrote:
>
> > The code quality of this package cannot be independantly verified as
> > source is not available and  using this driver, which works at very
> > low levels of the OS, can compromise the stability of the NetBSD OS
> > for which each and every NetBSD developer has worked so hard.
>
> so don't use it.
>
> > I suggest that announcements of this kind may not be allowed on the
> > list
>
> it upsets you that commercial products are available for our open-source
> OS?  if you want open-source PVRITY you know where to find the GPL
> camp...