[TriLUG] without better support from the vendors...

[Andrew C. Oliver]

>I did, back in September.  It looks like there are quite a few new 
>entries.  It also looks that RH7.2 works _much_ better that RH7.1
>did (which I tried without success).  Maybe I'll try 7.2.

+1  I was considering migrating to something else until 7.2.  

I really only run one heavy 3d-like video app....the screensaver.  It
seems that under the 2.4.7 kernel (that comes with 7.2) my video card, a
voodoo3, likes to cause a freeze up.  Later kernels do not suffer from
this.  Because I'm running short on disk space I just disabled all of
the 3d screensavers instead of upgrading.  I know this to be a problem
because I had upgraded to 2.4.7 when I was running 7.1.  Anyhow I don't
know specifically why or what, how, etc, just that it seems to happen
and I didn't have that problem with 2.4.9.  

All in all these video card issues are not uncommon to other operating
systems.  Windows NT 3.5.1 had similar problems with video drivers as
did OS/2.  Microsoft's solution to the problem may in fact not be such a
bad idea, they started certifying a few drivers/vendors.  Because of
their considerable marketshare and resources they expanded this rather
quickly.  (And you run into a chicken and the egg issue...you need
better video support to get more graphically oriented users, but you
need more graphically oriented users to persuade the video driver mfgs
to improve it)

It seems to me the linux distribution vendors already have programs like
this.  It seems reasonable that they enter into agreements with a select
few of the vendors and assist driver development, certify the drivers
and expand from there.  Personally I'd accept a driver with a sealed
binary.  Ideally they'd open it completely, but it seems reasonable to
keep it closed for a time, then once the next generation video card
comes out, release the old code.  Although i consider myself an active
supporter of open source software and technologies, I can understand how
neck and neck hardware competitors (such as video cards mfgs) might not
want to disclose the source to their video cards.

At the meeting the other night, I said something on the order like "Why
cant they release the source except for the patented memory routines"
and someone commented that the driver wouldn't work very well unless you
could address its memory.  What I meant was, the source except for that
part, and yes it wouldn't work, but it seems to me someone could write
an open algorythm for doing the same thing to replace the patented
part.  Perhaps I misunderstood the issue.

In any case, I have found, though I'm not much of a gamer, that Linux
has made a perfectly acceptable desktop for me.  Setup can be a bit of a
pain, but I don't have to do it too often.  (You can't really start out
with Redhat 5.0 and just upgrade each release to 7.2 and expect to never
have to start over configuration-wise, especially when you upgrade the
hardware too...not that you can with any other OS either, nor do I
really expect it).  

I have to say I like 7.2.  I will soon buy another disk drive and
upgrade the kernel (I have better luck with the things I do using the
new memory management that everyone complained about), but for now I
don't have any serious complaints with Linux as a desktop.  Of course
I'd probably complain if I used KDE instead of Ximian Gnome *g*.


-Andy


>> http://www.math.uakron.edu/~chad/linux_t23.html (Redhat, but info on
X)

>I had a link to this, I think from a Linux/Thinkpad mailling list I
>found...but it 
>was one of the e-mails that I lost :( 




>Thanks for it. 
>Chris


-- 
www.superlinksoftware.com
www.sourceforge.net/projects/poi - port of Excel format to java
http://developer.java.sun.com/developer/bugParade/bugs/4487555.html 
			- fix java generics!


The avalanche has already started. It is too late for the pebbles to
vote.
-Ambassador Kosh