Higer Exposure of OpenGL product

OpenGL
1

I know some of you aren’t going to like the proposal.

OpenGL needs more exposure I think it’s clear to most of you that OpenGL is underexposed and it getting less and less popular on the internet which means less popularity overall.

OpenGL (blue), DirectX (red)

Many developers use OpenGL but there isn’t really a mainstream presence.

I think it’s time for SGI to boost OpenGL as a free alternative to DirectX. Sounds backwards but that’s how it is in 2007.

OpenGL is perfect for indie developers because of its syntax and cross-platform functionality so why not market it like that?

XNA is very popular but it’s slow, massive and quite frankly redundant (IMO). Yet, people stop using DirectX and switch to XNA because it easier to use. A wrapper around an API.

I’m just throwing out ideas here and I think you know where I am headed with this. OpenGL is already user friendly (syntax) yet it’s not so friendly coming to actually implementing it on your platform of choice.

The Open Source and Freeware libraries are usually buggy, no longer being developed or simply too big.

So I’m suggesting two things: improving popularity and usability. Two things that can’t be bad.

I’d like to know what you think about this since I sincerely feel that DirectX 9 and 10 have seriously over crowded graphics development and Windows is becoming the platform.
Everyone knows this is not good; I might be wrong.

Eddy

2

I think it’s time for SGI to boost OpenGL
I think you’re out of touch with recent developments.

Look into the khronos group and it’s activities & do a bit of catching up.

3

I think OpenGL ES is gaining popularity as the mobile 3D solution. Is Kronos responsible for that success? OpenGL has the advantage there in that Microsoft has never (to my limited knowledge) played in that field (3D mobile).

OpenGL is underrepresented on the Windows platform because MS can provide a LOT of support for D3D developers. nVidia/AMD have to support both GL and D3D developers so as a whole the support for GL (on Windows) ends up being less than D3D.

It’s difficult the stress advantages of GL over D3D except with availability of geometry shaders on WinXP using GL (unavailable with D3D9). Kronos (or some entity) would have to milk that difference to boost GL presence. Unfortunately, software/hardware development cycles probably can’t create a pervasive need for geometry shaders before XP reaches EOL, thus eliminated a possible advantage.

Something I’m sure would help would be the release of the next Newsletter. :wink:

4

Hi all,

I thought that SGI was still performing local OpenGL seminars etc. Plus they still own all the rights to OpenGL if I’m not mistaken.

MS can provide a LOT of support for D3D developers.
I’m an MSDN subscriber with all the benefits, there’s no “direct” support for DirectX and XNA like there is for other products. The only support there is are the forums. Plus if you’re going to use your limited amount of reports and online concierge posts, who would waste it on D3D?

Unfortunately, software/hardware development cycles probably can’t create a pervasive need for geometry shaders before XP reaches EOL, thus eliminated a possible advantage.
Which is next year if I’m correct. Something that would really be an advantage would be a simple cross-platform windowing framework – not 3rd party, but by OpenGL – for C++ and maybe even C# developers to be able to create more programs without having to explore many many open-source options which are often incomplete or “dead”.

I also think that stressing the “Open” part of OpenGL is very important and making people aware of the consequences it has on applications. Look at the massive campaigns Mozilla Inc. ran for Firefox; they really paid off.

Eddy

5

OpenGL is not “open” in the same way as Open Source. OpenGL is open because the process for creating the spec is “open”. That is, anyone can participate (a fee or registration may be required).

Further, Firefox isn’t where it is because it is Open Source. It is where it is because IE6 was a huge security hole, as well as missing basic features of modern browsers. Microsoft fell asleep at the wheel and Mozilla stepped in.

Granted, Microsoft is falling asleep again, with the DX10 for Vista only thing. It isn’t quite as bad as leaving IE6 alone for 5 years, but it does give GL 3.0 an opening.

Getting us a spec for GL 3.0 would be the start of taking advantage of that opening.

6

OpenGL is open because the process for creating the spec is “open”.
That’s what I was aiming at since the DX team doesn’t have an option for regular community users to put in their $0.02.

Firefox isn’t where it is because it is Open Source […] Microsoft fell asleep at the wheel and Mozilla stepped in.
I didn’t say it was because of open source, but because of their campaigns such as: The full page ad in the New York Times, the Spread Firefox program and Google’s input through the AdSense campaigns.

The next paragraph might be opinionated: There’s a nice amount of money involved if you look at the amount of corporations that are listed as Promoting Members of the Khronos Group Inc, shouldn’t some of that cash be used for high level marketing? That’s $340,000.00 from Promoting Members alone on a yearly basis. Add on top of that income from: other memberships, conformance testing, research grants and corporate interest.

Getting us a spec for GL 3.0 would be the start of taking advantage of that opening.
But will anyone besides the already 3D savvy community jump on it? I think it’s time for some organizational action, that’s all.

7

Which is next year if I’m correct.
Um, what? XP reach end of life next year?

Even if Vista weren’t underperforming, you could expect to see many XP machines in the wild well through 2010. And Microsoft will still support it with patches and security fixes well past then.

shouldn’t some of that cash be used for high level marketing?
And who exactly do you intend to market to?

Firefox markets to people who have a computer. Ordinary users. Ads in magazines matter to them.

GL isn’t something that can get marketed to the public because they don’t use it. Indeed, the people that would be in the best position to help GL increase it’s userbase are developers.

Magazine ads don’t matter to developers; they’re not going to fundamentally change how their program’s rendering works because of a magazine ad. PR spin is meaningless to them. What convinces them are cold, solid facts.

Furthermore, $300,000 doesn’t go very far in terms of advertising.

But will anyone besides the already 3D savvy community jump on it?
Without a spec, there is nothing to jump on. And, as I pointed out, it is the start, not the end.

What OpenGL needs far more than a marketing blitz is to rebuild the trust relationship between IHVs and developers. Many developers opt to use D3D because it is more stable, with more consistent performance. They don’t trust ATi and nVidia; with good reason. ATi’s GL drivers are crap, and nVidia’s continue to have glslang compile bugs and needless slowness years after the spec shipped.

IHVs need to be absolutely dedicated to making quality, performant GL 3.0 drivers. And they need to not merely say it, but do it. 6 months (maximum) after the spec is released, developers should be able to have functioning, performant GL 3.0 drivers.

This will be of far greater value than a marketing campaign. After all, no matter how good Firefox’s marketing was, nobody would have stuck with it if it was crap.

8

I think that just about covers it. Any more business?

9

Well, I agree with Eddy… OpenGL needs a boost… where is our OpenGL 3! Is October…
I think OpenGL 3 gonna be very popular, because D3D10 is Vista-only and we all know that Vista sux!

Btw Eddy, how did you print that nice google graph?

10

MS also collects feedback from IHVs and ISV via the Graphics Advisory Board, so it is not just them specing DX out, though they have the final word. Pretty much similiar to the ARB.

Chris Donahue: David mentioned working with the IHVs and ISVs…there’s a graphics advisory board that’s made up of the big graphics brains around the industry. David and a lot of the guys on the DirectX team and the Graphics Platform Unit group spend time talking to, to find out, you know, five years from now, where are things going to be? And that stuff they take and turn into actual API code and ends up not only in the API but also in hardware. It’s a really collaborate effort between Microsoft, the ISVs that actually consume the API and make applications, and the hardware guys.
Source

11

Btw Eddy, how did you print that nice google graph?
It’s a direct image link to the Google Trends (google it) image. It’s also updated every day, so tomorrow the picture is different than today.

I understand everyone’s position on this matter and I empathize. I wasn’t referring to $300K as the only source of marketing monies, Khronos is a corporation, as is SGI; they have other sources of income.

OpenGL is an important (can I emphasize it enough?) factor in 3D Graphics development, if OpenGL dies there will not be anything besides DirectX -proprietary- and Mesa which is a software renderer.

It’s a really collaborate effort between Microsoft, the ISVs that actually consume the API and make applications, and the hardware guys.
That should be:
It’s a really collaborate effort between Microsoft, the software vendors that make enough money for Microsoft, and only the major hardware guys.

I am not against Microsoft, I actually pay to get their latest software and tools through MSDN but I fear Microsoft’s standpoint on Graphics Development is at an extremely high level (XNA, .NET et al) and restricting to only a certain demographic (.NET developers and Indie Hobbyists).

Side note: If you’re an MSDN user you’ll notice how many times they ask you “How large is your company?” or “How much profit did your company make last year?” If you make less than a certain amount don’t expect to be able to give any feedback to Microsoft.

Having an Open platform is important. Knowledge of its existence and its usability: more important.

12

GL3.1 will pretty much kill dx10 dead. Games developers will use it to make their games look as good on XP (the only significant OS) as they do on Vista, and if the drivers are stable enough, I see no reason why they would maintain their dx10 path.
MS will regret making a sacrificial lamb of dx10.

13

if OpenGL dies there will not be anything besides DirectX

GL3.1 will pretty much kill dx10 dead.
OpenGL is not going to die. Neither is DX10. From the simple fact that they play to different markets it won’t happen. I agree that there should be some more visibility for GL in the Windows market and the coming out of GL3.0 will definitely help that.

I see no reason why they would maintain their dx10 path.
If they have already committed to a DX10 path it would be potentially more difficult to rewrite it for a GL3.1 path, especially seeing as how GL3.0 isn’t even out yet, much less drivers, GL3.1 has got a long way to go. (I still am hoping they call it something other than 3.1. From a branding standpoint it makes sense to indicate how different it is from 3.0.)

14

knackered, you are joking right? Or is this just a dream of you?

15

First games that will actually use GL 3.x will not appear before two-three years from now. And that only if IHVs will deliver good drivers soon. Sorry knackered, I don’t see this happening.

16

I will have to agree on one point knackered made which states that the benefit of having DX10 features on Windows XP will be a good selling point for the not so distant future.

Yet: if Service Pack 1 for Vista comes out and the mass of XP users steps over to Vista, what’s the point? Then you’ll have a native API (DX10) competing with an introduced API. There has to be something besides “DX10 features on Vista” for the Win32 platform which DX doesn’t have.

IHVs may support additional features that will come out in the future but that has several implications:

  • GL Feature X is not a part of a standardized feature-set
  • GL Feature X may only be supported by IHV A
  • GL Feature X came out long after it came out on DX.

I think it’s coming to a point where there’s less and less innovation by IHVs, ISVs and end-users; to an extent that the APIs and Hardware implementations are now in a stagnant state.

Meaning that: yes there is somewhat an improvement in the visuals if you use a G80 chip and DX10 but is that all? There is no real speed improvement comparing speeds for DX9 graphics on DX9 architecture and DX10 graphics on DX10 architecture. And this is with the promise of Unified architecture and other graphics-load-balancing implementations already available.

What’s the real big feature that graphics APIs have introduced over the last 4-5 years? Nothing, if you really consider it. Not even DX10 which is now so highly regarded and praised by Microsoft and IHVs as a sales tactic.

Bad analogies such at this:
(Xbox Official Site: Consoles, Games, and Community | Xbox)
and this:
(http://www.winmatrix.com/forums/index.php?showtopic=9550)
have suckered in gamers and other multimedia enthusiasts by comparing low poly vs hi-poly and SM 4.0 shaders vs regular vertex based texturing.

PS, I love how Microsoft used their own screenshot (Halo) to signify obsolete technology.

17

First games that will actually use GL 3.x will not appear before two-three years from now.
StarCraft 2. It’s going to have a Mac release, and Mac only uses OpenGL. And there’s already been some suggestion that the ARB has consulted Blizzard (or vice-versa) during the development of GL 3, so there’s clearly some desire on their part to use it.

18

If GL3.1 is stable and consistent across vendors, it will be used in preference to DX10 because it will be a single, simple, efficient, capable, extendible API that’s available across every single operating system in use today. Frankly, it would be a no-brainer.

19

(Note I use DX to mean D3D due to laziness)

There has to be something besides “DX10 features on Vista” for the Win32 platform which DX doesn’t have.
Some consider portability/design to be features enough. What you’re thinking of is the graphics architecture which then gets exposed by the API. It just so happens that DX hasn’t had (doesn’t have?) an extension mechanism to always support the latest architecture. DX will always be upgraded to support the latest features and so it will always come down to API preference. Hopefully portability and API stability are attractive features for developers (they would be for me).

GL Feature X is not a part of a standardized feature-set
It depends what you mean by “GL Feature”. Part of the core spec? An extension? For an IHV to put out a GL driver it must conform to the GL spec. Also, IHVs have every incentive to support their newest hardware via extensions or supporting the newest API (whether it be DX or GL).

GL Feature X came out long after it came out on DX.
I’m pretty sure (someone else could verify) that GL tends to have access to the newest hardware features (via extensions) before DX, as DX with new features tends to be an entire version rev.

What’s the real big feature that graphics APIs have introduced over the last 4-5 years?
Exposure to the newest hardware programmability features. Vertex/fragment shaders have been around for a while now but how about geometry shaders? Maybe you missed GL2.0 and GL2.1? I’m unclear why you dismiss this activity. What would you prefer, exactly?

20

Originally posted by knackered:
If GL3.1 is stable and consistent across vendors, it will be used in preference to DX10 because it will be a single, simple, efficient, capable, extendible API that’s available across every single operating system in use today. Frankly, it would be a no-brainer.
A good API is a first step but as a game developer I want more than this. The secret of success is support and infrastructure. Direct3D offers me PIX, PerfHUD and FXComposer for free. Additional I got a reference implantation for quick checks if my code or the driver is wrong.