Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 25 Apr 2003 20:42:42 -0700
From:      Terry Lambert <tlambert2@mindspring.com>
To:        "Gary W. Swearingen" <swear@attbi.com>
Cc:        freebsd-chat@freebsd.org
Subject:   Re: Code layout and debugging time
Message-ID:  <3EAA0032.3CE3AF41@mindspring.com>
References:  <20030422132906.GB64101@dogma.freebsd-uk.eu.org> <444r4qmp6n.fsf@be-well.ilk.org> <20030422172549.GA65023@dogma.freebsd-uk.eu.org> <20030425044935.GG81840@dogma.freebsd-uk.eu.org> <3qel3q9k1r.l3q@localhost.localdomain>

next in thread | previous in thread | raw e-mail | index | archive | help
"Gary W. Swearingen" wrote:
> And, though you apparently disagree with me, the documentation shouldn't
> publicly humiliate the hardware manufacturer, as the "rl" manpage does.
> A private message to the manufacturer would be more seemly.  If I were
> RealTek, I'd withhold ALL documentation and communication with FreeBSD
> people until they stopped their anti-advertising campain against my
> company.

I don't know about that.  If someone builds bogus hardware, you
should probably call them on it.

I remember when Adaptec came out with their "HIM" layer to keep
people from implementing plug-compatible hardware that worked
with Adaptec drivers, and all the OpenSource people were out in
the cold, until they rewrote the sequencer code from scratch.
There are still some Adaptec RAID controllers for which the RAID
sequencer code has never been rewritten, and so don't support
RAID.

Likewise, I remember when Diamond Multimedia had a couple of
revisions of their video cards, and because they didn't table
drive the INT 10 ROM implementation, and wanted to be able to
do ECNs by reprogramming their PAL and their ROM, without
changing the card design otherwise, and so they didn't work
with XFree86 at all.  All because they didn't table-drive the
PAL inputs from a documented table that was stored after a
known ROM signature (which would have allowed Open Source card
independent drivers for the "Viper" and other cards, without
risking your hardware from bad PAL inputs).

Likewise, Winmodems.

Calling people on bad designs is almost a sacred duty.  8-).

I do know that I tend to read the drivers when trying to pick
between two pieces of hardware that nominally fulfill the same
role.


> I filed PR 31271 on this matter and it got multiple agreement
> from doc-ers and got "fixed", but it seems that somebody's vicious
> streak ran too deep and the nastiness was mostly restored -- and it's
> much worse in the driver source.  And if RealTek has improved, as you
> say, those comments certainly ought to be improved too.

The comments are, well, irascible is probably the best word;
still, being blunt is not all bad.

The new cards, which use a different driver, have improved.
The old cards still have all the problems they are documented
as having.  I'd prefer it if I were less likely to have an old
lemon unloaded on me because the new hardware unfairly raises
the brand reputation: "a rising tide floats all boats", as they
say; it's probably not worth buying an Edsel because you like
the new Mustang, or having "Car and Driver" retract what they
said about the Edsel, because they have nice things to say about
the new Mustang.  An Edsel is still an Edsel.

If you want a different analogy: Conexant makes real modems, and
Conexant makes Winmodems, but just because I like their real
modems, doesn't mean I'm going to buy anything with the name
Conexant on the box.

-- Terry



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3EAA0032.3CE3AF41>