On Wed, 15 Oct 2003, Alexander Shopov wrote:

> > The _only_ answer that matters is the
>> technical/scientific one.  End users opinions about how things 
>Technically and scientifically you are right and I agree with you, but 
>not everyone has the patience for the scientific side. I as sorry as you 
>are about this thing, but some magic some times is surely appreciated.

I don't disagree.

>> It doesn't help anything.  People will create rumours and spread 
>> them _always_ by the rules of human nature and the fact that the 
>> overwhelming majority of people don't understand deep technical 
>> issues in general.
>
>Maybe it will not help directly you but it will help me in several ways. 
>I will not have to dispel the myth about "the networking in X that slows 
>it down".

Point people to Keith Packard's whitepaper on keithp.com that 
discusses latency, etc.

>I will talk about RENDER or sth. else. And maybe people will
>stop pesting XFree86 developers to drop netwoking support. (Well
>they are sure to find sth else to unscientifically voice their
>opinion about, but let us make this one little step)

People can pester XFree86 developers or whoever they like to drop 
networking support.  That in itself makes no sense, since X11 
itself is a *NETWORKING PROTOCOL*.  Drop the network protocol, 
you drop the entire concept of X completely, and throw out all of 
the applications while you're at it.  XFree86 and any other X 
developers are simply not going to do that at all period, and 
they're unlikely to be convinced otherwise.  Other people are 
always free to go and implement their own end all be all 
windowing systems.  I doubt anyone who works on XFree86 or X in 
general will care either way if someone wants to implement an 
alternative to X.

What is funny however, is that any alternative to X, is more or 
less functionally useless until someone writes an X server for 
it for most general purpose computing.


>> Video gaming is a perfect example.  Playing video games is indeed 
>> possible in Linux using XFree86.  I would NOT "advocate" 
>> Linux/XFree86 to video gamers however, nor would I try to extoll 
>> the virtues of gaming in Linux with XFree86.  It does work, but 
>> it is not a push and click painless experience yet for the masses 
>> out there.
>
>Actually you ARE dropping several variables from the equation. Real life 
>example - we had recently in Bulgaria the following case: Microsoft told 
>computer gaming clubs that they could not use their bought and paid up 
>licenses for Windows 98 and let people "hire" the computers on a per 
>hour basis. Microsoft's view was that they needed Windows XP 
>Professional licenses. The clubs showed the letters they had with MS 
>partners from which they bought the licenses in which MS distributors 
>explicitly stated that Windows 98 is the necessary version that would 
>suffice (the letters were written maybe 2 years before Win XP was on the 
>market)
>What finally happened is that clubs got busted, non-compliance with 
>licences was found (as well as tax avoidance) and a club had more than 
>200 computers confiscated.
>So - for the end user perspective - gaming in XFree86 is not painless, 
>but for the point of perspective of game club manager - it is less 
>painful to have to pay your network administrators to make the thing 
>click than to have your machines confiscated.

Making video games run in Linux for use in a gaming club or 
internet cafe would IMHO be more an exercise in futility or at 
least a major headache.  It's nice to have the option 
nonetheless, but I wouldn't advocate Linux or XFree86 for that 
purpose as it isn't something that currently excels as a solution 
for that particular type of use.


>You are sayng that you need to make comparisons with things
>being equal.  You mean - hardware configuration and so on. But
>there are people for whom what matters is the cost - so hardware
>specs can be left aside. You can invest what you save from
>licensing the OS in more games or better hardware.

That isn't what I mean at all.  I mean you make comparisons with
the exact same machine running the exact same hardware for both
tests, and that the drivers for both operating systems are
configured to use the same features.

For example, running some major 3D game, and having the Windows
drivers configured to do 16x fullscreen antialiasing, and every
other 3D option cranked to the tits, while possibly having Kazaa
or something in the background with 10 people
downloading/uploading stuff, etc. and comparing that to a Linux
box with nothing running on it but the game, with no
antialiasing, and other 3D options not even available due to
driver limitations - isn't exactly comparing apples to apples.  
In order to do a fair test, one must configure both operating 
systems and drivers similarly and make sure that the sandbox 
being used to test both systems is as fair a comparison as 
possible.  Also ensuring things like Windows Prefetch service 
kicking in while the test is running, or Linux updatedb cronjob 
not kicking in need to be taken into account.

Benchmarks can be a useful thing to compare computer systems or
software with, but benchmarks can also be used intentionally to
highlight the best points of the system one wants to win, and
highlight the weak points of the other system.  It can also be 
done unintentionally by someone not realizing something is 
interfering or skewing their testing.

In short, benchmarks and similar tests are only one thing, and 
the information they provide is not 100% conclusive all around 
in a general sense.


>Will you help me show the magic in XFree86? The jaw dropping
>side of things?

Not sure exactly what you're asking here..  It takes a lot for my 
jaw to drop though.  I'm not sure XFree86, or any other computer 
software can make that happen though.  Ok, maybe the Halflife 2 
movie trailer comes close...  <grin>

;oP



-- 
Mike A. Harris

_______________________________________________
Devel mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/devel

Reply via email to