| |||||||||||||
MeeGo: the merger of Maemo and MoblinMeeGo: the merger of Maemo and MoblinPosted Feb 16, 2010 23:41 UTC (Tue) by ras (subscriber, #33059)Parent article: MeeGo: the merger of Maemo and Moblin
What is happening here isn't merging. Merging is Nokia marketing speak for the death of Linux Distribution they created. Put is simple terms, Nokia is ditching their own home grown distribution (Meamo) in favour of Intels Molbin. They are porting their Qt UI libraries and a few apps, but that is about it.
Maybe some of the reasons for this were revealed in this post:
http://talk.maemo.org/showthread.php?p=528755#post528755
Any death causes a lot of angst amongst those who were close to the deceased, and this is no exception. The points that came up on most on talk.maemo.org, were:
1. Will MeeGo run on the N900?
2. What will happen to to the maemo.org community? (Good question, as Nokia hasn't said. But would have to think it will be the same as meamo: what doesn't migrate over to Molbin will die.)
3. Dislike of the move from a Debian based system to Fedora.
The last question was the cause of most consternation, and probably for good reason. It seems likely Nokia will do their best to make the maemo.org community feel at home at MeeGo, and of course MeeGo will run on the N900. But Debian packaging is dead, primarily because RedHat managed to make rpm the preferred format for LSB. That means all the packaging work done by the maemo.org community is also now dead. It also means Debian's package repository, which is the largest on the planet, will be lost to the community. This is doubly sad because Debian has an official armel port and Fedora doesn't, and many of the current Debian ARM binary packages just worked on the N900.
In the end all this probably means very little to Nokia. They had a go at creating their own distribution. It failed, and so they are making a commercial decision to move on. The users at maemo.org represent very little in the way of sales, the users they will loose because the N900 was a phone running mini Debian distribution are vanishingly small. But many in the community at maemo.org has invested something worth more to them than money, and seeing this investment sacrificed on the hard alter of commercial reality is proving a bit hard to take right now.
(Log in to post comments)
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 0:05 UTC (Wed) by drag (subscriber, #31333) [Link] They should not be so down about it.Moblin and Maemo are very very similar and always have been. The differences between Fedora and Debian are actually pretty small. Sure you don't have the same ARMEL distribution support upstream, but Debian and Fedora all share the same exact code bases and similar versions for all but a tiny distro-specific fraction of the software. So the effort to make Fedora to actually support armeabi is going to be small. The biggest problems are probably going to get cross-compiling support for Fedora up to speed, but that is handled by GCC and other upstream projects just fine. Anyways. It's time for Fedora to start support ARM anyways. The ARM-linux user base has surpassed all other arches except for x86 and x86-64 for a couple years now. I would be surprised if it took much effort to get ARM officially supported if all of a sudden a new crop of users showed up. For a young project your probably better off with Fedora. Especially if you want to get packages supported by Fedora and become actual Fedora members. The time and effort it takes to become a Debian maintainer is very high; while it should be much easier to get involved in the Fedora project. You will find it's much easier to belong and support packages as well as get your changes in the base OS. People have tried to build off of Debian in the past and they all got kinda screwed over by Debian's inability to get releases out in a timely manner. (believe me, the link in your post is NOT unique to people trying to work with Debian) The only people that successfully done it is Ubuntu and that was because they were willing to do their own releases off of Sid rather then waiting around for Debian stable to surface. For every movement like this there are upsides and downsides. So while it is bad to move away from Debian; it is good also.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 0:15 UTC (Wed) by dlang (✭ supporter ✭, #313) [Link]
it's long past time for Fedora to support ARM, and several other architectures.
unfortunately it's more than just pointing a cross-compiler at your tree and letting it churn for a while. Debian has been doing the work, and a lot of testing. RedHat will now need to start doing this.
I also wonder how well it's going to work to base a phone distro (which may not be upgraded entirely for years) on a distro that only has a 12-month support window.
I've built small system images from both RedHat and Debian distros. It's _much_ easier to strip down the debian distro. The OLPC software fiasco had many causes, but I think that one of them was the problem of trying to balance between forking packages and living with the bloat.
In this case I suspect that after a year or so all that 'based on Fedora' will mean is that it uses rpm packages, but the software itself will probably have very little in common with the then-current Fedora distro.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 0:53 UTC (Wed) by drag (subscriber, #31333) [Link] unfortunately it's more than just pointing a cross-compiler at your tree and letting it churn for a while. Debian has been doing the work, and a lot of testing. RedHat will now need to start doing this.Well.. it kinda is for some developers. If your building a embedded system this has been very acceptable solution traditionally. It drives me batty, but for most devices you have a single application you need to support. Once you get that working on your OS then your goal is to never touch your OS again. It's done. So a one-time effort of compiling your own stuff and doing a bunch of hacks and work arounds is perfectly acceptable. Of course that is not really something useful for a long term project that wants to support lots of different stuff. It'll take some effort and it won't happen overnight, but most of the work is already done by Debian. Like I said before if you step back and look at the distros.. they are in their own little sandboxes, but they all use just about the same versions of software in the same ways with the same code base. There is relatively little variation. So most of the work that Debian has put into their ports will hugely benefit Fedora if they decide to go that direction. also wonder how well it's going to work to base a phone distro (which may not be upgraded entirely for years) on a distro that only has a 12-month support window. It probably does not matter a whole lot. Right now you'll need the latest and greatest QT4 version to get what is needed commercially. Is that in Debian Stable now? Nope. When is the next Debian stable going to come out and get that support? Not even Debian knows. It could be in 8 months, it could be in 2 years. What if you need to back port newer features? Any code you touch means your on your own. If Debian stuck to a time-based release then Nokia and other third party folks could coordinate and work with them. But the way it is now they can't. I know this is beating a dead horse, but that is just the reality of the situation. So while Debian may offer long term support for Stable.. it is pretty much worthless for anybody that wants to make a real product based on Debian. I've built small system images from both RedHat and Debian distros. It's _much_ easier to strip down the debian distro. The OLPC software fiasco had many causes, but I think that one of them was the problem of trying to balance between forking packages and living with the bloat. Ditto. Debian is indeed much easier to work with on small systems. Like I said before there are upsides and downsides. In the end it will probably be much easier to work with Fedora to sliming down their system and removing bloat then it is to get Debian to change their policies. In this case I suspect that after a year or so all that 'based on Fedora' will mean is that it uses rpm packages, but the software itself will probably have very little in common with the then- current Fedora distro. Yes. It would be unfortunate if the same thing that happenned to Maemo and Debian would happen to Fedora and Meego. Of course Fedora folks already have a Moblin version of their OS out. So it is not like the Moblin and Fedora folks have had the same problems that the Maemo and Debian folks have had.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 1:14 UTC (Wed) by dlang (✭ supporter ✭, #313) [Link]
I will also point out that RedHat used to support more architectures than they do now. They dropped them because they were more work to support than they were willing to expend. So far they have not shown indications that they want to start doing so again.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 5:38 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link]
You have no idea what you are talking about at all OLPC uses a stock
Fedora system with only the kernel being different and Fedora has ARM support for quite sometime now
http://fedoraproject.org/wiki/Architectures/ARM
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 6:10 UTC (Wed) by dlang (✭ supporter ✭, #313) [Link]
OLPC now uses a stock fedora process, but there have been many times when people attempted to trim bloat (very important on a machine with 1G flash and 256M ram) only to be told "but we have to do it this way because that's how it's done in Fedora"
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 6:17 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link]
Fedora is not a immutable system and OLPC folks have commit access to the
Fedora repository and can work with the maintainers to change things to be suitable for both as they have done that many times in the past to the point that nothing diverges except a few patches for the kernel and your quote needs a citation
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 16:21 UTC (Wed) by sbishop (guest, #33061) [Link]
"... and your quote needs a citation"
And your post needs punctuation.
I'm sorry. It rhymed. I couldn't help myself. :)
MeeGo: the merger of Maemo and Moblin Posted Feb 26, 2010 22:13 UTC (Fri) by oak (subscriber, #2786) [Link]
> http://fedoraproject.org/wiki/Architectures/ARM
"CPU and Architecture Target
Uh, that's not much of help on ARMv7 machines. Yes, the code runs, but it
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 0:25 UTC (Wed) by mbanck (subscriber, #9035) [Link]
Note that the Meego people (mostly Dirk Hohndel and Arjen van de Ven) insisted that meego
was its own distribution and not Fedora-derived at all.
They are apparently using OBS (Open Build System) for autobuilding, packaging and patching.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 1:00 UTC (Wed) by drag (subscriber, #31333) [Link]
That sucks. Somebody needs to slap them.
They are not building a OS for a specific product so that approach is self-defeating.
Off all the things the world needs.. it is NOT another Linux distribution.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 3:35 UTC (Wed) by xanni (subscriber, #361) [Link]
"NOT another Linux distribution."
Why not?
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 13:47 UTC (Wed) by drag (subscriber, #31333) [Link] Because it's a huge waste of f*king time and resources, that's why.They need to ask themselves: What is the entire point behind MeeGo? Is it to provide a compelling user experience with a slick UI, high quality mobile applications along with a solid set of application APIs along with a set of high quality development tools? If that is the goal then why the hell are you going to spend your time screwing around with Udev rules, GCC patches, Glibc configurations, custom packages, kernel configurations, just a nother half-assed shitty OS installer and crap like that? Waste of f-ing time and effort. Instead of concentrating on what is important your off screwing around with the same boring, error prone, tired crap that has been done over and over and over and over again. Dull and repetitive and it'll drive all your potential users and contributers away because nobody wants to deal with that crap. IT IS ALREADY DONE. I mean Fedora and Ubuntu already have Moblin OSes out there. They are compliant, have the user land, follow the specifications. It's FINISHED. Already f-ing done. So what MeeGo needs to do is work on the stuff people care about. They need to port over the Maemo 6 development environment and UI. They could be doing that TODAY. Get it running on Fedora or Ubuntu as quickly as possible and let users and application developers start doing what they do best... which does NOT include giving a shit about re-hashing a stupid installer or low-level Unix OS. I mean seriously look at: http://distrowatch.com/ _waste_of_time_. If you want to have some project were you say "Hey I have my Own OS, look how cool I am!!" that is fine by me.. but if you want to make a compelling user environment for a commercial product that will get widely used and open up new markets for Linux then starting off by making your own Distro is about the worst possible thing you could possibly do. Fedora has a shitload of developers. They have the attention of the Linux kernel development community, huge user base, high quality tools, high quality installers, high quality hardware support. Debugging infrastructure, the orginization is already setup.... All of this is already _DONE_. And they WANT you to use their system. They are screaming for people to use and improve their system. There is no hidden proprietary Goo. No Redhat holding them down. It's free and they want to work with you. What is the chances of MeeGo distro ever acheiving the same level of oginization and resources for it's own distro? NONE. ZERO ZILTCH NADDA. So MeeGo has the choice of using what is built and supported by others so they can start cranking away on creating the high quality UI, application, development tools and such that Linux has always needed to compete in the desktop or mobile environment commercially... or they can be another pointless half-assed distro screwing around with details that nobody cares about and have already been solved in a superior way by people with a hell of a lot more experience and resources.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 15:32 UTC (Wed) by HelloWorld (guest, #56129) [Link]
Because distros having multiple different distros was a bloody stupid idea to begin with. Fragmentation is a Bad Thing. Fragmentation in the UNIX market is what allowed the rise of Windows NT. It makes software development harder and it makes the deployment harder. A lot of documentation is useless, because it's written for a distro other than the one you are using. Tons of effort are wasted because different developers package the same packages over and over and over for different distros. And this also leads to stupid bugs like the debian openssl disaster or the recent bug in OpenSuse's screen locking thingy. Not to mention the fact that having fewer distros would make it a lot easier for users to pick one (Just watch the talk "The Paradox of Choice" by Barry Schwartz, you can find it via a search engine of your choice (no pun intended)).
My personal belief is that eliminating 99.9% of all distros would increase linux adoption massively. And while we're at it, we might as well kill gtk+ in order to give developers a uniform platform to the develop for (Qt, that is). Sometimes less really is more.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 16:21 UTC (Wed) by nix (subscriber, #2304) [Link]
And choice? Choice is bad.
Thank goodness there is no central authority with the ability to do what you propose. Having multiple distros is a wonderful way to experiment with different packaging and system administration methods, or to customize the OS for particular domains. Yes, a lot are derivative or boring: but a lot aren't.
MeeGo: the merger of Maemo and Moblin Posted Feb 18, 2010 21:52 UTC (Thu) by HelloWorld (guest, #56129) [Link] And choice? Choice is bad.So it is, if there is to much of it. And having literally thousands of distros is definitely too much. Having multiple distros is a wonderful way to experiment with different packaging and system administration methods, No, it's not. A uniform linux distribution for the desktop wouldn't stop anyone from developing, say, a new package manager or system administration tool. It would actually make things easier, because it would be possible to build and package a new tool so that *everyone* (almost) can take advantage of it. Imagine that: people could just use system-config-httpd to configure their apache web server. And if they don't like it, they remove it and install YaST instead. Having a uniform desktop linux distribution would actually give you *more* freedom, since you won't be constrained any longer by the selection of packages your distro has made for you. or to customize the OS for particular domains. Yes, this is why i spoke about eliminating 99.9% of all linux distributions and not about eliminating each one but one. I do see a point in having stuff like OpenWrt for embedded devices, or MeeGo for mobile devices. But having Ubuntu, OpenSuse, Mandriva, Fedora, Mint, MEPIS, Sidux etc. at the same time is utter nonsense.
The Free Software Mantra Posted Feb 17, 2010 3:48 UTC (Wed) by ncm (subscriber, #165) [Link]
Slap early, slap often.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 7:25 UTC (Wed) by wstephenson (subscriber, #14795) [Link]
I understant that they are using their own instance of the openSUSE Build
Service for building the packages and the distro. NB that the OBS tool comes from openSUSE but is not limited to building openSUSE distros.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 5:42 UTC (Wed) by arjan (subscriber, #36785) [Link]
I don't really understand what Fedora doing ARM has to do with Moblin or Maemo or Meego....
What Fedora does and what Moblin / Maemo / MeeGo do have nothing to do with eachother... at all.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 14:27 UTC (Wed) by drag (subscriber, #31333) [Link] If your talking about making a new OS... then yeah. Moblin/Meego/Moblin has nothing to do with Fedora.MeeGo should be doing absolutely nothing of the sort. They should be concentrating on stuff that is interesting and that is going to make it a commercial success. Trying to design a new operating system for cell phones and whatnot is exactly contrary to that that goal. They won't be able to do a good job and its a waste of resources. They should be concentrating on creating UI improvements, good applications, standardized APIs, and high quality development tools that can run on existing, easily customizable, and established operating systems; like Fedora or Debian.
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 2:32 UTC (Wed) by pabs (subscriber, #43278) [Link]
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 6:11 UTC (Wed) by ajk (subscriber, #6607) [Link]
> It also means Debian's package repository, which is the largest on the planet, will be lost to the community.
You ought to be able to run Debian in chroot no matter what the underlying system is, as long as it's built on top of the Linux kernel on a supported ISA and you can get root access. (As far as I can tell, this is how native Debian packages are used on Maemo even now.)
MeeGo: the merger of Maemo and Moblin Posted Feb 18, 2010 0:08 UTC (Thu) by mgedmin (subscriber, #34497) [Link]
Some people do use chroots, yes.
Other people take Debian source packages and recompile them in Maemo's SDK,
MeeGo: the merger of Maemo and Moblin Posted Feb 17, 2010 11:05 UTC (Wed) by Hanno (guest, #41730) [Link] Put is simple terms, Nokia is ditching their own home grown distribution (Meamo) in favour of Intels Molbin. They are porting their Qt UI libraries and a few apps, but that is about it.This is overly dramatic and even if you really want to see it that way, it looks like a good decision to join forces for something as dull as packaging and distributing software for a software stack so similar as Moblin and Maemo.
|
|||||||||||||