XBMC Linux port questions and answers... - Printable Version
+- XBMC Community Forum (http://forum.xbmc.org)
+-- Forum: Help and Support (/forumdisplay.php?fid=33)
+--- Forum: XBMC General Help and Support (/forumdisplay.php?fid=111)
+---- Forum: Linux and Live support (/forumdisplay.php?fid=52)
+---- Thread: XBMC Linux port questions and answers... (/showthread.php?tid=26097)
- jonb2 - 2007-05-29 21:44
No your missing the point. The same linux that ran on a 386 back in 1995 will run on a brand new just built last week from not yet released reference design new hardware pc. You don't pick a niche product to be the only hardware you support you pick the most common.
XBMC on linux is not relying on hardware, it is relying on linux. They are not creating an operating system they are creating an application, and application that needs certain video and audio requirements. So you choose the hardware that is most popular and least problematic.
- ultrabrutal - 2007-05-29 21:56
jonb2, most common meaning any pc hardware in your case. show me one motherboard and graphicscard from today that you can buy in 5 years. heck even in 6 months. xbmc will not be limited - it's opensource. you can run it on your 1995 pc for all I care.
xbmc rely on linux and linux rely on hardware - hence xbmc rely on hardware too. say you are having trouble with your xbmc pc because you are using some odd graphicscard. where will you seek help? nowhere! you have to figure it all out on yourself. most people are end users who just want it to work stable and as intendeed. you will not get this with your bamboo pc.
- mace - 2007-05-29 23:29
The only way to get HW stability with major vendors like HP and Dell is to buy their corp modells. One of the reasons for their higher pricetags is the longer lifespan of those modells.
And stability is the holy grail. It doesn't matter how fast it runs, on wich HW or what features it has if it isn't rock stable. In my opinion it has to be as stable as a vcr or a dvd recorder. In other words better than most current set top boxes.
That in turn requiers a fairly small amount of HW support in order to keep testing at a reasonable level because you can't just say that one graphics adapter works. You have to say that this adapter works with this mainboard with this bios rev.
- raid517 - 2007-05-29 23:38
Sorry for the comment abuse. But may I just bow and scrape in both utter grattitude and worship for those who have taken this bold decision?
I have wanted this to happen ever since XBMC was first released - and have begged and pleaded for it on many, many occasions.
You guys rock - and XBMC rocks like no other Media Center app. on the planet and now (or in the near future perhaps) I can fulfil my dream of building a small dedicated XBMC machine and of finally retiring my almost antique and venerable old XBox1.
It also resolves another dilemma in my mind - which is the requirement to compile XBMC (an open source and wholly honest application) using an illegal (and in most cases stolen) SDK. (An SDK that is also almost undoubtedly becoming rapidly out of date, particularly in light of the ever increasing demands of HDTV and new media types, both current and those yet to be envisaged).
But now XBMC has a future - and it deserves a future. It does not deserve to die as the XBox1 user base continues on it's ever more rapid decline.
So I take my hat off to you guys and I say thanks. Thanks for all the years that have passed - and thanks once more for hopefully all of the years to come.
Some things make life that little more bearable and that little bit more enjoyable - and in my view XBMC is certainly up there among the best of these.
- BLKMGK - 2007-05-30 02:59
I too have been awaiting news like this and just spotted the recruitment announcement, this is wonderful! I've read with interest the comments on hardware support. One of the things I've moaned for with Myth has been for those guys to pick a solid danged HW platform and support the heck out of it the way XBMC has been supported on XBOX hardware. I think that limiting support to a specific couple of platforms is a great way to go.
The aTV is an excellent CHEAP platform if it's powerful enough, the Mini or the upcoming Mini also good, the 360 would be nice but umm yeah okay maybe someday. Pick a solid patform for this, I'll buy it - simple as that. I have multiple XBOX1 (give them as gifts with XBMC on it), and two 360s - one with old firmware. If you pick the aTV I'll own one or more of those too The solid support is worth it for both end user and developers IMO. The idea that you somehow support 360 peripherals is also pretty awesome, if it's successful that will be terrific. I wouldn't necessarily be looking to choose a platform someone is going to have in their closet but something that works and isn't too tough on the developers. The Apple hardware, and I own none of it, looks pretty good thus far....
Good Luck guys, I for one am certainly rooting for you and when things are far enough along will happily purchase hardware and help test!
- anywonder - 2007-05-30 03:28
Gamester17 Wrote:The nice thing about Apple TV is that unlike the old Xbox (which uses customized chipset and GPU) Apple simply took a standard Mobile Intel Centrino 915PM Express chipset (featuring Intel ICH7 7.1 High Definition Audio) and a standard NVIDIA GeForce Go 7300 GPU (graphic processor chip), making the Apple TV basically just a laptop computer with a proprietary operating-system and without a built-in keyboard/trackpad and display that normally is attached to laptop computer. So we can just use the normal proprietary binary device drivers that NVIDIA provides for Linux, (or we can choose to use the open source drivers if we clike, but I doubt those are as good as the closed source one NVIDIA provides, at least not yet, though I'm not sure what the license sais about distributing the proprietary binary device drivers from NVIDIA on Live CD).
Hey Gamester, there's a project (like r300) to reverse engineer nVidia's proprietary drivers for all their cards (so they don't bother ppl when they, on schedule, drop support for older cards), called Nouveau. I believe they're working on Xv support and then someone is doing Xvmc support for the Google Summer of Code. Chart on functionality. I believe other distros, like Kororaa, have been forced to stop shipping w/ nVidia drivers due to kernel developers' copyrights and GPL on the kernel.
Intel's open source graphics are working on supporting Xvmc, apprently in the Xorg 7.3 git tree.
ATI's proprietary drivers, well, I'm sure you know the problems with them. Don't even support AIGLX and the reverse engineered drivers don't keep up fast enough because ATI's hardware is buggy.
- anywonder - 2007-05-30 03:30
*wanted to add that they're not "reverse engineering" in the technical sense nVidia's drivers, but watching what they do when renouveau runs opengl on the binary drivers, and then implementing it in Nouveau.
- bmfrosty - 2007-05-30 04:52
I just thought that I'd throw in my two cents. I've been using XBMC for a number of years now, and about two months ago, I gave up on ever being able to play 720p H.264 on my XBOX, so I broke down and bought a PC and have been masochistically using mediaportal since. I am so happy to hear that the XBMC team is working on a linux port.
Given all this, I've been thinking about what a future platform for XBMC could be. None of the current game systems are very appropriate for the job, and while I like Apple platforms, and they would probably be a good place to start, I think that running with VIA EPIA would be a better idea.
First and foremost, the EPIA platforms are all about being open source for all of their features, including video acceleration. This is limited to MPEG2 and MPEG4 at the moment, but it would surprise me if they didn't support H.264 and VC-1 at 1080i/p by this time next year.
Secondly, VIA tends to keep their platforms around for a while, new revisions come around all the time, but they tend to be incremental changes that should be easy to adapt to.
Thirdly, the EPIA line is based around the philosophy that low power quiet systems are better. This is important for anyone who is primarily using this as a HTPC solution.
Via already has a very nice HTPC board available in the Mini-ITX that's built for HTPC usage - the EPIA SP:
I fully expect that within the year they'll have a version going at 2ghz that has H.264 acceleration on board.
As for the rest of the hardware, I've been thinking that while it's very expensive to get custom motherboards, there are plenty of manufacturers that could build a cheap plastic case with built-in power supply and IR Receiver and bundled remote that they could sell for a profit at $50 or $60. I know that we're nowhere near where we'd need to be in order to start talking to manufacturers. There are plenty in china, and interestingly enough I've got a co-worker that goes to china on a regular basis and deals with these guys.
Lastly, while having a hard drive and optical drive is very sexy, for most of my purposes, I don't need either. It seems to me that once this gets going, it could probably fit on a bootable 512 meg usb flash drive, OS included. This would likely be more than enough for those of us looking for a solution for network streaming.
I hope that this helps when you're making your final decisions on supported platforms, and as an aside, when you're get to that point, I may be able to round up a couple of boards as test platforms, any admin should be able to look up my email address.
Also, it might be worth your whiles to try and get in touch with the EPIA team about what you're doing. They may have some interest in it as well.
- jonb2 - 2007-05-30 05:43
I wonder if some of you understand the point I was making. So I want to clarify a few things, once ported to linux this will run on ten thousand hardware configurations, no ifs and or buts. You could load it on red storm, a blade server, 4 year old dell or a brand new custom built machine.
What we are talking about is the hardware configuration that will be recommended and that this site, the developers, and forum users will try to provided -free- support for. Now what many of you are saying is that you only want to provide support for a specific mac. Think about that for a minute. Why would want to avoid supporting the vast majority of your user base? More then half the people that will use this will using it on an AMD or Intel cpu with an ATI, Nvidia, or Intel onboards graphic's and realtek, or sound blaster audio that they may well already have, but you only want to support a pair of Mac's. That just doesn't make sense.
On another note virtualization was mentioned several post back and think this is an excellent tool for this sort of thing.
- bmfrosty - 2007-05-30 06:01
I see what you're saying, but I think you need to think about what a daunting task support can be. Of course users are going to be using the software on many different platforms, and I'm sure many developers will be experimenting with multiple different hardware configurations, and by default any bugs found there will probably end up being fixed. The fact remains that there is a need for there to be a core platform for support purposes.
I really don't think you need to worry about it too much. This type of project is likely to attract a lot of help over the next few months, most of whom will be planning to make sure it works on hardware that they personally like.