[kryptonyte]

Ditto... actually.

NS

I see what you did there....cute.

I've been keeping up with the thread and ask the dev a few questions. XBMC-NMT (The name he used) is going through testing and they are working really hard to optimize the GUI to look smoother and cleaner. It has been discussed that lighter skins such as Alaska run better, though still not a smooth as on ION systems.

Playback is going smooth from what the testers have reported and most major bugs have been found a fixed so a release should be coming up here within a month or two. Given this started March 1, I would say it is progressing fairly well.

[tennisbgc]

Hats off to the dev! - but as it stands it makes my Xbox look like a quad core.... and it doesnt actually connect to anything in the clip I dare say they will sort it thou.

Well the pch only has a 300 mhz cpu and I believe the xbox is 700 or so.

[waldo22]

Here's a few things I haven't heard answered yet.

I read the whole thread there, and a few people asked, but no one ever answered...

1) How is he rendering the GUI? from my (limited) understanding, the PCH boxes don't support OpenGL, or any 3D acceleration API for that matter. Will this require some kind of custom 2D skin? (he said regular Confluence works, albeit sluggishly...)

2) How is he accessing the proprietary Sigma HW Accelerated decoding hardware in the NMT/PCH? Without this, I don't see how he could play even SD MPEG-2 video, much less HD, with a 300Mhz MIPS chip, so I guess he's figured it out somehow

3) If he is using the Sigma proprietary drivers and such, what are the legal ramifications? Will we have to be careful about distributing binaries a la xbox?

4) I guess this means XBMC now runs on MIPS? Did he use the ARM port at all as guidance for this? Did he single-handedly port to MIPS by himself in 2 months? McGeagh (and others) have been working on ARM for quite a while...

I'd be interested in the Dev's opinions on this, and EJP's (if he's reading this).

[jmarshall]

We don't know the answers to any of the above. Obviously we've been in contact with EJP but we know nothing more than you guys at this point.

One presumes that the work done on getting XBMC running on i386, PPC, ARM etc. makes it easier to work on other platforms though.

Cheers,
Jonathan

[ubuntuf4n]

I wonder if it will be able to handle aeon one day , would be cool.

however, first of all a full working system should be more than enough...

[waldo22]

We don't know the answers to any of the above. Obviously we've been in contact with EJP but we know nothing more than you guys at this point

Thanks Jonathan,

I didn't mean to imply that anyone owes us these answers (especially anyone from Team XBMC ) I'm just wild with curiosity.

It seems like an awful lot of crazy things had to happen to make this work.

If he's ported to MIPS, gotten some form of 3D acceleration working, and gotten the Sigma HW Acceleration working all in 2 months, then that's pretty damned impressive.

@ejp, if you're reading this, drop us some dark hints.

EDIT:

More answers (and videos) will be revealed as development furthers. Technical aspects are currently shared only between me and the beta testers.

If it's any consolation, I can say that the method of how rendering and playback is being done has already been speculated in this thread.

Guess details will be revealed in due time...

-Wes

[saratoga]

We don't know the answers to any of the above. Obviously we've been in contact with EJP but we know nothing more than you guys at this point.

Unless I'm missing it they don't provide the source for those XBMC binaries they're passing around. How nice of them.

[natethomas]

Unless I'm missing it they don't provide the source for those XBMC binaries they're passing around. How nice of them.

GPL doesn't require releasing code until you release the software. So as long as they keep the program within a small development group and don't release it to the public, they're fine.

[topfs2]

GPL doesn't require releasing code until you release the software. So as long as they keep the program within a small development group and don't release it to the public, they're fine.

To be more accurate. GPL states that anyone getting the binary must have the code available to them, and the anyone that have the code have the GPL rights to distrobute.

This means thats its perfectly correct to develop a GPL app in the dark and release it whenever they see fit, so long as when they release the binary to the public they get the code aswell. And so long as the people that have gotten the binary have gotten the code, they are alright.

[natethomas]

To be more accurate. GPL states that anyone getting the binary must have the code available to them, and the anyone that have the code have the GPL rights to distrobute.

This means thats its perfectly correct to develop a GPL app in the dark and release it whenever they see fit, so long as when they release the binary to the public they get the code aswell. And so long as the people that have gotten the binary have gotten the code, they are alright.

I stand expounded upon.

[CrashX]

Looks like ejb has updated with XBMC Playback of SD video. He has promissed to update with new information every tuesday.

Does the GPL allow me to distribute a modified or beta version under a nondisclosure agreement?

No. The GPL says that your modified versions must carry all the freedoms stated in the GPL. Thus, anyone who receives a copy of your version from you has the right to redistribute copies (modified or not) of that version. You may not distribute any version of the work on a more restrictive basis.

[darkscout]

No. The GPL says that your modified versions must carry all the freedoms stated in the GPL. Thus, anyone who receives a copy of your version from you has the right to redistribute copies (modified or not) of that version. You may not distribute any version of the work on a more restrictive basis.

And they reserve the right to remove everyone that distributes a copy of the binary and source from beta testing.

It's a gentleman's agreement. "Hey, I want some more people to beta test this, just don't send it out." Legally he can't do anything, but it'd be a dick move. Look how many rabid new people there are that want all the newest features but wonder why X feature of SVN isn't working. Imagine the flood of noise they'd get if someone said "Hey guys, look we have XBMC for Popcorn!". Heck even the AppleTV thread was filled with "Are we there yet posts".

[saratoga]

Look how many rabid new people there are that want all the newest features but wonder why X feature of SVN isn't working. Imagine the flood of noise they'd get if someone said "Hey guys, look we have XBMC for Popcorn!". Heck even the AppleTV thread was filled with "Are we there yet posts".

I didn't mean to imply that he had to give the source out to the whole world immediately (looking back I see that my phrasing was awful), but rather that what you're suggesting is kind of an ugly way to do things. IME it works much better to give binaries to a pool of testers, but the source to everyone. That way you keep out the flood of duplicate bug reports from people who aren't involved enough to contribute, but still allow potential new developers into the project, not to mention anyone upstream to provide feedback before you spend too much time implementing things the wrong way.

With rockbox (an embedded operating system/firmware) we have a publicly available SVN server for each new port, but don't give out a compiled bootloader until its ready for average users. This way to contribute bug reports you also have to be able to read and understand directions, and then compile something. Cuts down on the noise 100 fold but still usually lets us pick up a few new outside developers for each major port prior to release.

[natethomas]

With rockbox (an embedded operating system/firmware) we have a publicly available SVN server for each new port, but don't give out a compiled bootloader until its ready for average users. This way to contribute bug reports you also have to be able to read and understand directions, and then compile something. Cuts down on the noise 100 fold but still usually lets us pick up a few new outside developers for each major port prior to release.

That doesn't work as well with XBMC. We haven't been officially compiling builds since March, yet somehow (ahem) people have been getting their hands on builds using other methods. Every post about getting a tv-server to work is a post that we are officially not answering, because we said we wouldn't answer anything post those march builds and tv-server addons only came out after the fact.

With any luck, though, it won't be long now.

[kryptonyte]

After reading through the thread I'm actually beginning to doubt the authenticity of this. Too much has gone unanswered by ejp and the entire project seems to be hidden by some kind of ominous secrecy.

All he releases are "videos" with no further information about developments or anything else. All questions about how he managed to get software rendering of OpenGL to run smoothly on 300mHz MIPS processor, how he got smooth playback on that processor using ffmpeg and if he isn't using ffmpeg what he is using for playback?

None of this is touched on in either of the two threads he has on the NMT site or here. I know if I had managed to port XBMC to MIPS in only a couple months time, I would be damn proud and want to share how I did it. It all seems fairly fishy.

If this does turn out to be a real working port of XBMC to NMT then more power to him, but for now, I'm not getting my hopes up until I see/hear more details on how this is working. I think it was someone on the NMT forums that suggested this could all be a VNC client to a Windows/Linux/MacOS box running XBMC and commands from the Popcorn Hour remote sent through VNC as recognizable commands.