Why graphics drivers are proprietary

Marko Vojinovic vvmarko at gmail.com
Wed Oct 3 13:53:31 UTC 2012


On Wednesday, 3. October 2012. 2.55.25 jdow wrote:
> On 2012/10/03 02:01, Marko Vojinovic wrote:
> > How about government funding? There is a tried&tested scenario used for
> > some
> > time now all over the world, say in science. For example:
>
> Then you get what the government says you will want not what you do want.
> We saw that in Soviet Russia as a very glaring example.

That depends on how government is organized. In many countries the government 
does not decide how the funding is actually used and which R&D projects are 
financed. Those decisions are left to expert teams or peer review committees or 
such institutions. People who get to evaluate project proposals are typically 
the people who were voted by the community to be in those positions. It's 
called democracy. ;-)

You shouldn't judge the whole idea based on one lousy implementation that 
happened in Soviet Russia.

> > * You need money, and you have some skill to do something better than
> > others. * You apply for a research&development project; if you have a
> > good idea, you get a grant.
> 
> Is this how you'd start Google, Twitter, or Facebook? My, how quaint.

Why not? If there is a serious need for an advanced search engine, and if you 
think can do better than the existing facilities, go ahead and make a 
proposal. If it makes sense, why do you think you would not get a grant?

Twitter and Facebook are really social/amusement/entertainment stuff, and they 
started out with almost no funding at all, so I don't see any relevance here. 
Neither of them is made on any groundbreaking new technology that needed 
funding to develop.

Remember that the issue that started this topic was the performance of 
graphics hardware and drivers. If you have a promising idea to construct 
superior graphics hardware, or better drivers for the existing hardware, I'm 
sure many people will want to listen.

> > * You use your knowledge to do something creative and useful. You share
> > the
> > results of your work with everyone else (you're being paid by taxpayer
> > money, so this is fair).
> 
> Am I paid MORE if I produce something creative, whether or not the
> government wants it?

Your paycheck depends on the size of the grant money. If your proposal is 
important/useful/successful, you'll get a bigger grant. It's quite naturally 
organized.

> How about if customers want it and the government
> does not, especially if the government does not want it?

Look, the same system exists and works even in USA, in academia. The 
government are the people who are democratically elected to govern. If you are 
not satisfied with their decisions, vote for a different set of people in the 
next elections.

Besides, the government gets to decide only how to cut the global state budget 
into pieces and what to use those pieces for in general, not in particular. 
For example, if 0.5% of state budget is decided to go into "computer hardware 
development", the government itself is not so interested on which projects 
exactly this money gets spent. These things are decided downstream by 
committees where expert people are sitting. If you have a successful history 
of past projects in graphics hardware development, you may get to be in such a 
committee, and decide the faith of other people's project proposals, based on 
your past expertise.

This is being done in academic research all over the world for quite some time 
now, and works quite well, on the average. All serious countries in the world 
have implemented this (and no, it has nothing to do with Soviet Communism). 
I'm just saying that this model can be extended to other enterprises, which 
are atm working under the "free market" paradigm, and suffering because of 
global competitive behavior present in that paradigm. The nVidia/ATI example 
is typical.

> > * You apply for the next R&D project, and the next, and the next... You
> > build reputation according to your performance, and in time get bigger
> > grants, bigger money, etc.
> 
> You only get funding for what the government has declared the citizens
> want. Can you imagine an iPhone designed by a government? My imagination
> is not that strong.

Can you imagine the graphics card designed by the team of experts funded by 
the government? I can, and I bet that it would be far superior in quality, 
since the team that gets to design it does not care about cutting corners or 
profit margins or trade secrets or such stuff. They would concentrate on making 
a high-quality product, which would be of completely open design, so that 
others can build on it and later create an even better piece of hardware.

> > This scenario is not optimized to make most money, but to make best
> > quality
> > products. Others can build on your work and your knowledge, and you can
> > build on theirs. It's a model which promotes cooperation instead of
> > competition.
>
> No, sir, it is optimized to produce what the commissars declare you will
> build. And commissars seem to have a lamentable disconnect with the people
> they own.

This is completely false. The "commissars" are people like you and me, which 
are given the power to steer the R&D processes because they are trusted by the 
people to use that power wisely.

> >> If I know how to do something that people really want and can live
> >> comfortably on what I can earn doing this, by what right does anybody
> >> come in and tell me I have to share my know how with all and sundry
> >> so that I'm stuck cold and hungry because I can no longer earn money
> >> performing my unique service? That is the foundation if the concept of
> >> intellectual property.
> > 
> > Umm, no, what you are describing is called a "trade secret". And it is
> > completely ok, even necessary, to have trade secrects in the free market
> > scenario (as opposed to the government-funded R&D scenario that I
> > described
> > above, where trade secrets are disfavored and disfunctional).
> 
> Do you realize that you are contradicting your screed above? Video driver
> software IS trade secret information, Kemo Sabe.

Where do you see a contradiction? I know that video drivers are a trade 
secret, that's why they are closed-source in the first place. I am just saying 
that trade secrets are not exactly the same thing as "intellectual property" 
(but only a subset of it).
 
Best, :-)
Marko




More information about the users mailing list