[Open-graphics] Vision statement still not satisfying

[Bart van Leeuwen]

On Mon, 6 Dec 2004 13:47:16 +0100, Nicolas Capens wrote
> H Bart,
> 
> >> Our target market isn't only developers, is it?
> >
> > Without developers no Linux, so this is in fact also relevant to linux 
> > users
> > (and those of other open soucre operating systems for that matter)
> 
> Sure, but do you expect to sell the cards to be sold to end users 
> too? I really hope so, and that should be the real goal we're after.

Of course, but I also keep in mind what is realistic with the given resources
in the current market.

> And ultimately I don't see the strict need for open specifications 
> and documentation to reach that goal.
> 
> It can be a requirement of one possible path to reach that goal. If 
> no other approaches are even taken under consideration, fine with me...

Please read why the project got started in the first place. Read the
introduction page to the mailinglist, Timothies posting on lkm.org, on my
weblog etc etc. Open specifications and documentation is a goal in itself
also. You may not need them, others do.

If I want to go another way I could go to a hardware producer that makes nice
closed source drivers, those exist and there is no need to start an open
graphics project if that is all you need.

> 
> > When interested in games currently I will use an NVIDIA card. As long as 
> > their
> > support remains as good as it is I'll keep doing that most likely. That 
> > said,
> > playing games is a small part of what I do with computers. Testing new
> > software and operating systems is soemthign I do a lot more. For that 
> > nvidia
> > cards are not very usefull unless 2d support is enough.
> 
> I certainly understand that you do a lot more with computers. But 
> frankly, that's not what the end user is interested in. They want to 
> run 3D applications. And like you said support for that is good. For 
> the little problems remaining, again from the end user's perspective,
>  developing new hardware might not seem like the most optimal 
> approach to solve a software problem. Who wants to buy a second 
> graphics card if they got a really good one already, which costed 
> less and has much higher performance?

You only think about one type of enduser. None of my customers requires high
performance 3d. They require a stable desktop environment with enough hardware
acceleration to ensure it is responsive. They ened it to run office and
business applications. None of those have a need for 3d performance that
approaches a mx440, let alone that of a high-end 3d card.

The same applies to the huge majority of business users.

> 
> You have to understand that many of these end users won't have the 
> technical knowledge to understand the difficulties at the software 
> side. With the current price/performance/quality ratios, marketing 
> of this product is going to be very hard. Not a lot of people will 
> look at the low-level problems it solves.

Nope, which is why I believe that this card needs to cater to some niche that
can be served well within the current limitations. Again, high performance 3d
is almost surely noit a part of that given the constraints of the hardware.

> 
> > The specs as they are would be very usefull for making a home theatre or
> > digital video recorder/player device if entertainment must be in there. 
> > No, it
> > won't make a good gamign device if that is what you mean.
> 
> Home theatre system are nice, yes, but how are you going to compete 
> on that market? Besides, it's 2D we're talking about now, support is 
> adequate for these applications if I understood correctly.

No you do nto udnerstand correctly. Yes, there is 2d support for most existign
cards, it is however noit optimal, and many easy to achieve features are
difficult to access or unavailabel due to lack of documentation. THis
specifically applies to things like tv out support, hardware yuv-rgb
conversion and scaling, and other features relevant for this.

> 
> > Again, there is more in this world then gaming, and a modern desktop
> > environment uses a fair bit of 3d graphics as well. 3d hardware 
> > acceleration
> > is usefull even if it doesn't allow ou to play doom 3.
> 
> I know there's more than gaming, absolutely. I'm not a gamer myself. 
> I'd just like to know what applications, the end user's interest,
>  will be able to run better on Open Graphics hardware.
> 
> > Expect to? nope. It is not impossible, but it is extremely unlikely. I 
> > doubt
> > this is a problem at all.
> 
> Good.
> 
> It is impossible though...
> 
> > For all I can tell we are indeed not too clear yet about where this niche 
> > is
> > tho some areas have been indentified and are mentioned also on the about 
> > page
> > below the vision statement.
> 
> It's nice to finally see a bit of confirmation that the vision 
> statement, or at least the targeted market, is not clear yet.

I am starting to wonder... I quote from the page:

    * The target market for this product is currently identified, but not
limited to: Enthusiasts, Universities, Linux OEM Integrators, Graphics
specialists and other niche markets.

    * The tarket market for this product is NOT to be a direct competitor to
the mainstream graphics market. (i.e. Nvidia, ATI, Matrox)

I think it is at least clear there that the target market has only been
indentified in part, and also that competing with nvidia and ati is not really
part of the game.

Did you really read this? or are we arguing in a typical slashdot manner
without rtfa?

> I truely don't expect an answer right away, I just hope more people 
> will realize this is important for the success of the project.

I'd understand that if you did point out somethign utterly lacking. SO far
imho you managed to confirm that there are limitations to what the project can
do, and that the assumption that we shouldn't even try competing with high-end
3d graphics at least initially is a good one. THat however was already stated
before we started this discussion.

> 
> > That said, maybe you want to suggest a text that words this all better. I
> > wrote the initial version of whats there now, but English is not my native
> > language, so well, it might not be the best way to state whats there.
> 
> Waarom denk je dat Engels mijn moedertaal wel zou zijn? ;-) Anyway,
>  it's not me who is going to carry out the project. I'm a poor 
> student and I'm trying to make money writing software, not wasting 
> my time on writing drivers. That being said, I -will- support this 
> project in every other way I can. If you want me to perform tests 
> with swShader, share knowledge about algorithms, say my opinion 
> about design decisions, that's fine. I've just given up trying to 
> understand the goals of this project, so please don't expect me to 
> define any.

You seem to understand the limitations fairly well. What makes it so hard to
understand the goals given those limitations? The difficult bit is to make a
match with as big a market as possible within those constrains. If you really
feel like supporting the project, it would imho be a good idea to keep the
limitations in mind and help find a target market that can be well matched
despite those limitations.

If you believe there is an entirely different approach possible, well, go try
it. Many of the things you have suggested have been tried more then once however.

Bart.
--
Created with Open WebMail at http://www.bartsplace.net/
Read my weblog at http://soapbox.bartsplace.net/