Possibly offtopic : Binary only driver

[Mike Hearn]

Hm, for some reason your posts aren't showing up on gmane. I don't know
if it's just slow or there's something more strange going on, so I'll
reply to all the points here in one email.

> > So what? It still involves debugging programs with binary components.
> 
> If you'd ever done any work of that nature you wouldnt be saying "so what"

Please don't attack me, go for the arguments. That way you won't make
wrong assumptions based on limited knowledge.

I've been working on Wine for about a year and a half now, and these
days it's my job. There are apparently no interesting open source Win32
apps because the programs people want to run are always proprietary.

So yes I _HAVE_ done a ton of work of this nature, and yes I still say
"so what". It's not easy, it's not always fun, but there you go: that's
life. Sometimes you have to debug problems with other peoples binary
code loaded into the process. This does involve quite tricky things like
debugging race conditions in closed source DLLs, but it can't be
avoided. It's either that or Linux isn't useful to a particular class of
people who need app XYZ.

I am wishing the kernel developers would change their attitudes with
full knowledge of what debugging binary only code is like. I do it all
the time.

> You've also clearly never spent time talking to Nvidia about the issues as
> they see them.

Clearly I haven't, but you have, and like I said last thing I heard from
you was that you couldn't justify them open sourcing their drivers. Have
you changed your mind, or did I misinterpret what you originally said?

Are you saying that they could GPL their drivers and not lose any
advantage at all over other proprietary gfx card vendors?

> No. Nvidia debug the Nvidia stuff nobody else.

Well apart from this:

http://diary.codemonkey.org.uk/index.php?month=5&year=2004

and there are a few posts on the xorg list where people take a look at
problems with the nvidia drivers.

> I do X development, I must disagree. Most X screwups result in either a
> return to text mode and a logged signal catch, or at least the ability to
> log back in and clean up/reboot nicely. Very important on server systems.

Yes, important on server systems. I was talking about desktops. An X
server crash kills every client connected to it (unless they install
specific panic handlers). That is the same as needing a reboot for
desktop users, except it's a bit faster to log in than reboot. But they
still lost their session and any unsaved work.

> Wrong. Simple rephrase   "The availability of plans of the house is useless
> except to builders". Now that should be obviously garbage to anyone. 

That's why I said "primarily useful". Obviously open source code is good
for society in general, that's why we're all here! My point was that
it's not so directly useful to non-programmers that they'll say "Oh well
I'll not play 3D games anymore" en-masse. Not going to happen.

> If you want any support with certain binary only modules (in limited cases)
> buy an enterprise product which is priced appropriately for the problems it
> causes

I don't want tech support, I want a stable ABI driver developers can
work to. If they screw up and my system crashes because of their driver
I'll run crying to them: somehow this system trundles along in Windows-
land, warts and all.

Finally:

> You are confused. They 4K stack merely showed they were already broken.

Of course in this instance "broken" is defined as "uses more stack than
the amount we've decided isn't broken". I understand the reasoning
behind the choice of 4k so I won't claim it's arbitrary, even though the
stacks could be larger. Windows uses 12k stacks does it not?

I guess we have different definitions of broken. 

thanks, your input is appreciated -mike

-- 
Mike Hearn <mike at navi.cx>