To: None <email@example.com>
From: Miles Nordin <carton@Ivy.NET>
Date: 12/08/2001 19:02:38
en0nn> Allowing the announcement of 4Front Tech's new sound driver
I agree with Herb, so far as that a transparent announcement policy
lends legitimacy to, well, to Herb and NetBSD's PR norms, by not
making us seem publicly like religious zealots.
At the same time, I think we ought to actually be religious zealots.
For example, it would be a mistake not to mention that 4Front's code
has, some five years ago when I crossed paths with them, been of
astonishingly low quality compared to the norms for sound drivers
across all vendor Unix platforms, and has shown an insulting
disrespect for host-OS internals like ISA PnP, and their
partial-forking of Linux's sound driver framework caused more harm
than good in the long run. Some Jesuits might want to hesitate
mentioning 4Front's selfish chainsaw-coding practices since 4Front is
``supporting NetBSD'' with their port.
Anyway, things have probably changed a bit at 4Front now that they
have more than one employee. It's too bad they kept the old name.
herb> I understand that kernel code can not be audited for quality
I think the point was that binary-only code cannot be peer-reviewed.
Kernel code certainly can.
A secondary point was that it is less important to audit application
code like Microsoft Word for NetBSD, since such code does not tarnish
NetBSD's reputation for quality by making kernels crash.
A transparent procedure to address this would be to peer-review
products, even those under NDA, before announcing them.
Unfortunately, not enough announcers will pay for this.
Certainly there is some precedent for preserving a product's good name
by branding accessories and plugin modules as Playstation-approved,
OSS conformant, Netware ready, Pure Java, Confirmed by NetBSD; or even
passing laws to indirectly prohibit your products' interoperability
with partners who you deem incompetent, who do their work outside
Japanese borders, or who don't pay you.
But i am guessing 4Front is not interested in paying NetBSD priests to
peer-review their code under NDA. The cheaper structural alternative
to our queasyness, of me issuing Fatwah that ``4Front sucks'' over
some moderated medium, is not transparent, or maybe even correct. so
it is relegated to the backwaters of advocacy while 4Front's
disgusting press release goes on announce. Disturbing, but correct.
There is no precedent for ``announcing'' that your competitors or
enemies are worthless scum although, of course, they are.
Linux is a complex GCC test-case, not a serious attempt at a kernel.
-- Eamon de Valera