On Wed, 7 Jul 2010 06:22:56 -0600, Nicola Mfb <[email protected]> wrote: > On Wed, Jul 7, 2010 at 1:32 AM, Dirk Hohndel <[email protected]> wrote: > > On Tue, 6 Jul 2010 17:06:57 -0600, Nicola Mfb <[email protected]> wrote: > [...] > >> * parts of sdk not running on ati/nvidia boards is community driven? > > > > The SDK uses interfaces that not all open source drivers expose. No one > > is stopping people from fixing that (i.e., fix the ATI and NVIDIA > > drivers). But please don't tell us that we need to use the least common > > denominator in order to be "open". > > I'm not expert on the specific issue, may you elaborate more? are > those interfaces intel stuff or something widely discussed and not > implemented by lazy opensource developers? are there implemented in > closed source drivers?
They are all implemented in open soruce drivers and have been part of the upstream X.Org / Mesa packages for quite a while. I believe that the open source ATI drivers implement most (if not all) of what's needed, I don't think the open source NVIDIA drivers do. I don't have access to a system with ATI or NVIDIA graphics (huge surprise, I guess), so I can't really check. > >> * not official support of meego on some hardware is community driven? > > > > Again, the default builds that we provide are optimized for Atom - I > > don't think there's anything wrong with that. It's fairly straight > > forward to build for other platforms if you need that, but I think it is > > not a reasonable request that we shouldn't optimize for our platform. > > Here I'm speaking about *official* support, If I buy the next Intel > atom netbook with meego, and a my code snippet will segfault due to a > gcc bug, I'm quite sure that I can open a bug on your official issue > tracker and you'll support the resolution even if it requires a gcc > patch and upstream contribution (funding). That would be my expectation, too. > I'm not sure that will happen if I will use a community supported > meego build on unsupported hardware, but that will happen for sure on > debian, gentoo, openembedded etc. > That's the reason I'm asking what do you mean for multi platform > revolution open new gnu/linux os, and what is your place in the > distribution scene. The MeeGo open source project needs to decide how to support different platforms. One of the contributors to the project, Intel, has decided to spend significant resources on supporting its own platforms as well as the overall project (and make no mistake - 95+% of what we do is platform agnostic and works on AMD, ARM, anyone). I find it very weird that we are somehow expected to promise and fund full support of competitors' platforms - when did that become the norm for being open? I repeat, if the MeeGo community wants full support for some AMD or MIPS or PPC platform then no one is stopping this. Get those companies to donate build resources (as those are currently one of the limiting factors) and find the engineers to work on these projects. Intel has donated build resources and is funding engineers for its platforms. > >> * "basic" software components on handhelds devices (gui telephony, > >> connectivity, audio routing, 3d etc.) will be part of meego and > >> opened? > > > > Not exactly sure what you mean here - to the best of my knowledge we'll > > have a complete stack, though not all pieces are in place, yet. > > Are you saying that any mobile vendor that provides the right kernel > may ship meego full featured cellular/tablets without writing a single > line of userland software and capable of running all the meego > compliants applications (for example extreme 3d games, or the > legendary closed ovi map free gps navigation rpm too?) in the same > exact way as the next nokia meego device? I have no idea what "extreme 3D games" are - and I don't speak for Nokia and their proprietary apps. What I said is that we are contributing a complete open source stack that to the best of my knowledge covers all parts that are necessary to create a netbook, tablet or phone running MeeGo. There are the usual caveats about specific hardware that may not have open source drivers (PVR graphics and some modems come to mind), but the rest of the software stack /should/ run on any hardware. Someone please correct me if I'm wrong. > > I really think that we are providing an open revolution. I don't think > > anyone is being held back to build this on other platforms or to add > > missing features to drivers that don't have them. Isn't that the idea of > > an open source project? > > True, just demonstrate that open here is a wide term (and respect the > community "pressing") with facts! on the atom side you are right (and > may perform better with support for other cpu), on the arm side just > support definitively meego for the n900. I may be excused for not trying to speak for the ARM side. > You want users, testers, community developers, apps buyers, so where > is it the hidden problem to migrate the ready maemo community? > Why the difference between atom line that supports existing hardware > and the arm one that seems to target (officially) upcoming secret > devices? I may be excused for not trying to speak for Nokia. > >> Many time you have "please respect peoples working under NDA" or in > >> the worst case the silence. > > > > Really? I follow this mailing list quite actively and can't remember > > having seen this phrase very often. > > Sorry, I'm not speaking about this list, just collecting info from > maemo forums, mailing lists, irc, and so on, and AFAIK maemo+moblin = > meego, so I think I'm speaking to the same guys. I'm have not done a full search of maemo forums, mailing lists and irc. My apologies. > >> As a random ml reader, I'm very curious to know how much openness may > >> be compatible with profit companies and business models. > > > > We are doing this in the open BECAUSE of we want to be successful and > > make money. I've spent the last 9 years at Intel getting people to > > understand that this is in our best business interest, and I think I > > have succeeded so far. > > Yes! I agree totally! and understand (hope) the meego strategy and its > winning possibilities, just want to note that it's quite natural by > Intel, and *seems* less by Nokia (that has to fight in a more > aggressive market, and has to "boom" with next high end devices). > Anyway I'm not able to see meego totally devoted to pure opensource os > development, in some way you have to sell your hardware and not give a > nice gift to your competitors, right? In reality we are giving a HUGE gift to competitors here. But we are being attacked for that huge gift not being 100% complete. It feels like an alternative reality at times. Intel has put hundred of man years of development into this project by now - and all of that is available in open source. > That's the reason I think a true open development and governance is > not fully possible, and a lot of internal resources will be opened > only *after* the release of the first meego device (I'd like to read a > big "you are wrong here" ;) ) No, you are not wrong at all. This is the "big reveal" discussion all over again. We are trying very hard to make as many components openly available as possible, but there will continue to be product related components that are not. But that has little to do with open development and governance in the long run. These things will be related to product specific user experience and potential value add features. MeeGo itself is open today (and the number of small pieces missing are more an issue of "it's not ready, yet" than of "we are holding things back"). Example: you can build a netbook image today. For any hardware that you'd like (some manual porting / driver work / optimizations may be required). /D -- Dirk Hohndel Intel Open Source Technology Center _______________________________________________ MeeGo-dev mailing list [email protected] http://lists.meego.com/listinfo/meego-dev
