Lets make a SDK 0.4 Linux nagging thread!

Hello again!

Now that the Samsung Gear VR cat is out of the bag* (cat not included), is there any chance we could get an update on the Linux support? My hope is that the complete silence on this front has been due to the Linux-oriented (and Android, by extension) source code potentially leaking the news of the collaboration. I don't know enough about OSX to know if there would have been a leak hazard there. :P

To quote your Kickstarter post:

It’s been an adventure keeping this project under-wraps for so long, and we’re thrilled to finally share it with you. Now that the word’s out, we’ll have more mobile VR news in the coming weeks, so stay tuned.

Are you thrilled to finally share Linux support information, too? Eh? :D

Optimistically but firmly prodding,
particlese

*See also:
https://www.kickstarter.com/projects/1523379957/oculus-rift-step-into-the-game/posts/973597 (publicly viewable)
http://www.samsung.com/ch/news/local/samsung-gear-vr (not yet available in a variety of English locales, despite the samsung.com link in the English Kickstarter update)
Edit: Just noticed the link on the Swiss press release to samsungmobilepress.com (not quite samsung.com)...
http://www.samsungmobilepress.com/2014/09/03/Gear-VR-1#Features (you may need to click the "features" tab to untruncate the press release)

"particlese" wrote:
My hope is that the complete silence on this front has been due to the Linux-oriented (and Android, by extension) source code potentially leaking the news of the collaboration.

I really doubt the source code for the android thing has anything to do with the linux SDK, because:

• they said they've created a new separate SDK for the GearVR and android, which makes sense, because the android windowing system has nothing to do with the X server used on regular GNU/Linux systems


• the code they had to write for GearVR was very hardware-specific kernel drivers to bypass all the intermediate android layers (java) and access low-lattency sensor info directly, which again obviously has nothing to do with the regular PC SDKs


• the GearVR doesn't even have use a positional tracking camera, so it's not like they took this chance to port the positional tracking code to linux with v4l or anything

All in all I think what they did for GearVR has very little bearing to anything else, and the people working on that sounded as a completely separate team lead by Carmack anyway.

Ah, nuts, you're right across the board, of course. I should have thought a bit harder before posting, but in my morning haze I got excited about the potential implications for Linux. :?

Untethered VR is obviously interesting, too, but I'm personally much more interested in using the desktop SDK on Linux even though I won't have my DK2 hardware for quite a while. I could look at the Windows SDK, but it's hard to get motivated in that direction when there hasn't be a peep about Linux support since... *searches thread for "cybereality"* well, just under 6 weeks ago. Sure has seemed longer than that...bah, Internet Time.

And since I'm here in this "Linux nagging thread": <nag>Dude, Oculus/cybereality. A "yes, we're still working on it, but I can't say anything more" or some sort of communication about the Linux SDK would be very much appreciated from many of us. Please? C'mon, you know you want to! <fud>Otherwise, we're left to assume Linux ineptitude, Zenimax lawyers, Facebook shenanigans, and/or SteamOS-induced Valve Time goodness are sealing your lips.</fud></nag>

The Linux SDK is being worked on, we haven't abandoned it. Everyone here is busy getting ready for Oculus Connect, so the bandwidth is spread pretty thin. Once the event is over, people on the team should have some time to dedicate to get this out the door. Sorry for the delay.

"cybereality" wrote:
The Linux SDK is being worked on, we haven't abandoned it. Everyone here is busy getting ready for Oculus Connect, so the bandwidth is spread pretty thin. Once the event is over, people on the team should have some time to dedicate to get this out the door. Sorry for the delay.

Perhaps you could talk to Valve about staff and such. And it couldn't hurt to get them involved either way... Steam OS/Linux might be the second largest platform the Rift can run when the consumer release comes to pass.

"cybereality" wrote:
The Linux SDK is being worked on, we haven't abandoned it. Everyone here is busy getting ready for Oculus Connect, so the bandwidth is spread pretty thin. Once the event is over, people on the team should have some time to dedicate to get this out the door. Sorry for the delay.

Thanks for the info :D

I'm curious, I recently answered your survey regarding what hardware/OS developers who own an OculusRift use.
Can you tell us how numbers are spread ? Or must it remain confidential ?
Did it clearly show that priority for unix platforms should remain lower than the rest ?
I was hoping steam boxes could propably shuffle priorities.

Thanks,

Thanks for checking in with the reassurance, cybereality, and sorry about the nag even though this is a nagging thread. :) I've been massively busy with a move, so Oculus Connect was completely off my radar. Keep concentrating to make it the best it can be, best of luck to all involved there, and I hope it bears all sorts of good fruit!

"cybereality" wrote:
The Linux SDK is being worked on, we haven't abandoned it. Everyone here is busy getting ready for Oculus Connect, so the bandwidth is spread pretty thin. Once the event is over, people on the team should have some time to dedicate to get this out the door. Sorry for the delay.

Thanks very much for the update. It's appreciated.

I guess it wasn't considered that anybody developing on Linux might have wanted to prepare something to demo at Oculus Connect?

It's a shame that Linux is such an afterthought for Oculus. As we've been discussing on this thread, it's actually the most suitable platform for the Rift.

The fundamental problem with focusing the Rift effort on the Windows platform instead of Linux (no, I'm not going to really argue that development should be Linux-centric) is that Windows is a terrible platform for this kind of thing, causing ridiculous grief by default until you write deep OS hackery to route around all the cruft. By comparison, Linux and X (and XRANDR / OpenGL) are well suited to adding extra displays with specific needs that are trivially independent of the primary displays. There's plenty of support for this, read:

A Trip Down the Graphics Pipeline
- which is specifically about the Oculus and the sharp contrast between the Win/X platforms -
http://doc-ok.org/?p=1057

Anyone here interested in software dev should read it immediately. There are also a few additional comments at:

https://news.ycombinator.com/item?id=8297729

Of course, if the aTDtGP is too optimistic, feel free to describe what it's leaving out. It's just one perspective.

My general impression is that the effort required to get the system to work right under windows may be impacting progress with other aspects of the Oculus. That doesn't mean the development for Windows isn't needed - it is - but it comes at a cost to the enabling of contributions from other developer communities. I also wonder if a Windows bias could lead to unfortunate choices is how to do abstraction around the X and OSX systems.

Anyway. I'd really like to see X, OSX, and Windows support to be treated as equals from an SDL prioritization perspective. A turning-point kind of app could come from any of those dev communities, a killer app might bridge *all* of them.

I've been away for a bit, so I had to look up what the Oculus Connect was.

I'm shocked that a social gathering has a higher priority than finishing a (advertised) promise. With the FB buyout, I'm sure there is much flexibility with "them" doing marketing, and letting the actual team code...but I guess not. :?: