OS2World OLD-STATIC-BACKUP Forum

WebSite Information => Article Discussions => Topic started by: Andi710 on 2011.05.17, 16:59:36

Title: About Snap Technology
Post by: Andi710 on 2011.05.17, 16:59:36
That is very interesting. Especially that they would license the source code for free.

Regarding expertise: Don't we have a graphics guru looking for bouties (demitrioussharpe)?

Regarding funding: I am willing to put EUR 500 on the table for a start and possibly more in case of tangible progress. And I imagine that many people would jump in as well.

Regarding scope: Focus on ONE desktop and ONE laptop chipset. Probably high-end at the moment - until it's implemented, it becomes mainstream. Implement all funtionality that makes sense (and is feasible) for OS/2 and eCS:
- fully accelerated 2D
- accelerated video
- all resolutions that the hardware provides
- SMP safe (enhancement over current SNAP, which seems not to be)
- multi-screen (possibly including a solution for laptops to use a second screen
- display orientation (albeit not on the fly as OS/2 does not support on the fly res change)

Out of scope (at least initially):
- accelerated 3D - only if someone had an idea what to do with it under OS/2
- multiple cards (crossfire etc.) - no use without massive graphics intensive games
- ... anything other that would delay a first release ;-)

Question is: could mensys 'own' the code and find a good project manager (like diver aka Sylvan) who can manage this as well as Qt and Java are managed. Then funding can also be raised. If it becomes another one of those zomby projects like flash, ACPI or various bounties at os2world.com, I don't think many people would be willing to invest (including me).

Cheers
Andreas
Title: Re: About Snap Technology
Post by: warpcafe on 2011.05.17, 18:11:58
Hi,

I might be mistaken, but... if I remember correctly, Mensys tried several times to buy the thing or license it but it didn't work out.
Or the price was insanely high? Not sure... but no, if memory serves, it wasn't even for the money but rather a "general" issue.

I will ask Roderick to put his 2 cents on the table... wait...

Other than that, yes, I am willing to collect all pending bounties and add some more if that gives me a fast 2D that supports two monitors with different resolutions at the same time. That's already all I would need.

Cheers,
Thomas
Title: Re: About Snap Technology
Post by: miturbide on 2011.05.17, 20:47:07
My personal comments had been posted on the Yahoo group thread (http://tech.groups.yahoo.com/group/eComStation/message/85802).

It is good to see that this is moving. But I don't consider the "just licensing" solution from Alt Richmond Inc. I prefer paying for an open source solution that will open the rights of licensing the software. Otherwise in the future you fall into legal voids and the software turns into abandomware when the money stops to flow.

Sorry, but I love to speculate. Nothing of this may happen but it is good to imagine the options.

What is the best scenario or the best options for me:
1.- Negotiate ($) to turn OS/2 SNAP source code open source and improve it from there.
2.- Negotiate ($) to turn PANORAMA source code open source  and improve it from there.
3.- Invest ($) of porting some 2D driver from other Open Source sources (Linux)

uhm.., the negotiation skills ($) are basic here ;)

But if there is going to be inversion from the community, the best thing is to invest in something open source that will not turn into abandomware and have legal voids when we lost the negotiation skills ($)  ;D
Title: Re: About Snap Technology
Post by: Mike on 2011.05.17, 21:27:14
hi,
agree with both statements from martin and warpcafe. First of all Alt Richmond showed no interest in the OS/2 part in the time mensys was asking them. The details about licensing are still mysterious because it means to share the core of the SNAP technology, thus preventing it from being open source and like martin said be dead again in the long run. We should better focus on something new and forget about panorama as a main graphics driver, its nice to have it as a fall back solution, but we want an easy nice working solution like SNAP was.
Title: Re: About Snap Technology
Post by: Andi710 on 2011.05.17, 23:31:52
The question is: What option is realistic at all? SNAP sources have specific code for OS/2, which a capable programmer should be able to compile and tweak for stability (SMP safety). Then it's all about adding the modules for oe or two additional chipset(s). The technology is proven - and still really cool considering what drivers are available on other platforms.

Open sourcing is out of the question for the guys at Alt Richmond and we cannot buy the entire company just to get OS/2 drivers done ;-) So we can dream of open sourcing the stuff, but it aint gonna happen. Unless those guys don't see any value in SNAP any more (won't happen in the near future).

Thus, if the license were free I would go this way and invest into a solution on current high end HW - this should carry us another 3-4 years. (The driver to the ATI X850 that my two LCDs are attached to is that age at least).
Title: Re: About Snap Technology
Post by: Blonde Guy on 2011.05.17, 23:34:38
Open source is preferable, and SNAP will never be open source.

An open source solution like GL/2 may come out. Suppose it was workable. To make a supported version would cost plenty. Assuming it cost about the same as the SNAP drivers, would you pay?
Title: Re: About Snap Technology
Post by: miturbide on 2011.05.18, 00:12:18
I try to see it in a long term. Why invest in future abandomware? 

We give the money, they do the driver and after 5 years the driver no longer works because there is new hardware and we are in the same situation.

If the solution is open source it can evolve with time without depending of one single software vendor.
Title: Re: About Snap Technology
Post by: Mike on 2011.05.18, 23:12:27
All about Alt Richmond was strange from the start. They had an emty website and showed no interest making a business case with the OS/2 version. Did they ever added new drivers to the SNAP core or sell this software to somebody else? To pay the license doesnt mean to have new drivers included...
Title: Re: About Snap Technology
Post by: Andi710 on 2011.05.19, 00:29:12
Quote from: Blonde Guy on 2011.05.17, 23:34:38
An open source solution like GL/2 may come out. Suppose it was workable. To make a supported version would cost plenty. Assuming it cost about the same as the SNAP drivers, would you pay?
Open source drivers are not impressive so far. Uniaud is OK, but not great, panorama does not allow for multiple screens and ACPI is just a disaster. I use OS/2, not LinuX, os2apic.psd, not ACPI, c-media driver that also works well on my SMP system, SNAP for two screens. Open source is just sane and CUPS (although I could also use a simple PCS or PS driver).
Open source is great for apps, that can be ported easily (e.g., based on Qt). But for drivers I am simply not convinced by the end products. I would certainly invest into something that works for the next 5 years rather than into some long term future dream that never materializes (like ACPI).
Title: Re: About Snap Technology
Post by: miturbide on 2011.05.19, 16:08:36
Hi Andy

Comparing Open Source with Linux it is not right. Linux is an example of Open Source software but it is not all that exists in OSS. Open Source is not Linux. Saying that a software is/should/must be open source does not mean that it is going to be Linux style.

Saying that a software is Open Source means that is released under any of OSI approved OSS approve licenses (http://www.opensource.org/licenses).

ACPI and Panorama are not open source AFAIK (please anybody feel free to correct me If I'm wrong).  UNIAUD is open source and have the benefit that can be further improved, while other audio drivers are just dead without any posibility to me enhanced.

Are the CMedia drivers you are using from Ruediger Ihle? because it says "This is a port of the open source C-Media PCI Linux audio driver"

I just want to show you there is more possibilities for the long term future if we invest in Open Source. If the developer quits, other developer can continue it. If the development company gets out of business other company can legally continue the project. That is the real long term benefit.

Please don't take it wrong, your point of view and arguments are also valid, I'm just expressing my way to see this SNAP possibility.
Title: Re: About Snap Technology
Post by: miturbide on 2011.05.19, 16:14:28
..also Andy. I don't disagreed with you that ACPI, Panorama, UNIAUD, and any other software requires more work, but this is complete separate thing if they are OSS software or not.  It depends on the amount of efforts and resources needed to be improved.

OSS does not mean it is good software by default, but it means it has the possibility to be improved against becoming abamdomware. I see it as a way to reduce risk for the user at the long term.
Title: Re: About Snap Technology
Post by: Mike on 2011.05.21, 19:12:16
Quote from: miturbide on 2011.05.19, 16:14:28
..also Andy. I don't disagreed with you that ACPI, Panorama, UNIAUD, and any other software requires more work, but this is complete separate thing if they are OSS software or not.  It depends on the amount of efforts and resources needed to be improved.
thats exactly te point, open source will have the same quality level when it is maintained like a closed source product owned by a company in case of development and defect support.
Uniaud is far from being perfect, but it serves basic sound support on most hardware now. We can be happy that IBM/Innotek made use of the ALSA core and implemented this technology for OS/2. The sblive and cmedia driver are outstanding drivers when it comes to features and stability however its not suitable for todays onboard sound devices, to develop a driver for each single sound chipset or a similar driver core like ALSA is not the way to go.
ACPI may not be important for old hardware but when it comes to new hardware it may be the only way to boot OS/2 on such hardware.
Title: Re: About Snap Technology
Post by: Andi710 on 2011.05.22, 15:04:42
Yep. I agree that open source is not to blame. Was shooting into the wrong direction. My apologies. However, I maintain that updating something that actually works (and was one of the tech highlights for OS/2 over the last decade) sounds more realistic than starting a new port effort based on some LinuX source and architecture. Given that I use some very old software components and they work very well, I'd rather invest my money into evolution than some long term dream.

I don't say 'don't do a new video driver model for eCS'. I would be excited if anyone succeeds!
I just wouldn't invest money into it.