We're sorry but this page doesn't work properly without JavaScript enabled. Please enable it to continue.
Feedback

Mobile distributions and upstream challenges

00:00

Formal Metadata

Title
Mobile distributions and upstream challenges
Title of Series
Number of Parts
97
Author
License
CC Attribution 2.0 Belgium:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Identifiers
Publisher
Release Date
Language

Content Metadata

Subject Area
Genre
5
15
Thumbnail
48:33
41
Thumbnail
35:21
47
48
Thumbnail
1:03:30
50
75
Thumbnail
50:56
94
BitDigital rights managementOpen source2 (number)Software bugMeeting/InterviewComputer animationLecture/Conference
Software bugOpen sourceSoftwareSoftware industryElement (mathematics)Computing platformOperator (mathematics)MiddlewareSelf-organizationDigital rights managementComputer animationLecture/Conference
View (database)Multiplication signComputer animationLecture/ConferenceMeeting/InterviewXML
Right angleBuildingLecture/ConferenceXML
Distribution (mathematics)Moment (mathematics)JSONXMLUMLLecture/Conference
BuildingComputer hardwareBitDistribution (mathematics)Military baseNumberComputing platformMobile WebComputer animationLecture/Conference
Source codeOpen sourceCASE <Informatik>CodeDistribution (mathematics)Computer hardwareBuildingPerspective (visual)BitSource codeComputing platformComputer animationMeeting/Interview
Computer hardwareOvalSoftware engineeringWhiteboardView (database)Point (geometry)Software developerOpen sourceComputing platformComputer hardwareOpen setReal numberWebsiteOperating systemOnline helpMultiplication signSoftwareWhiteboardOvalComputer animationLecture/Conference
Computer hardwareSoftware engineeringWhiteboardComputer hardwareOpen sourceDistribution (mathematics)Digital rights managementComputing platformTerm (mathematics)Computer animationLecture/Conference
Computing platformMechanism designTerm (mathematics)BitDistribution (mathematics)Lecture/ConferenceMeeting/Interview
Distribution (mathematics)Physical systemLecture/Conference
Distribution (mathematics)BuildingConfiguration spaceLecture/ConferenceMeeting/Interview
Operator (mathematics)SoftwareCASE <Informatik>Block (periodic table)BuildingDistribution (mathematics)AgreeablenessLecture/ConferenceMeeting/Interview
Lecture/Conference
Compilation albumQuicksortLecture/Conference
Distribution (mathematics)Computing platformAdditionLecture/ConferenceMeeting/Interview
Compilation albumMultiplication signConnectivity (graph theory)Computing platformOpen sourceLemma (mathematics)Computer animationLecture/Conference
QuicksortComputing platformLecture/Conference
Virtual realityDistribution (mathematics)Software developerCore dumpCartesian coordinate systemFreewareBitLecture/ConferenceMeeting/Interview
EmulationComputing platformBitSoftware bugOpen sourceOpen setLecture/Conference
EmulationDistribution (mathematics)Computing platformComputer hardware
Open setBitComputing platformMaxima and minimaLecture/ConferenceMeeting/Interview
Exclusive orEmulationMaxima and minimaComputing platformAndroid (robot)Line (geometry)Connectivity (graph theory)Open setSoftware developerStandard deviationBitDistribution (mathematics)Process (computing)NumberCartesian coordinate systemLecture/ConferenceMeeting/Interview
Software developerNumberComputing platformSource codeRaw image formatMeeting/Interview
Computing platformOpen setPhysical systemOperating systemBitSuite (music)Multiplication signSoftware developerMeeting/Interview
Open setStandard deviationComputing platformComputer configurationAxiom of choiceMUDRight angleMoment (mathematics)Point (geometry)Software developerLecture/ConferenceMeeting/Interview
CodeDisk read-and-write headView (database)Source codeSoftware developerComputing platformPatch (Unix)PressureOffice suitePerpetual motionWebsiteMedical imagingOpen setLecture/ConferenceMeeting/Interview
Patch (Unix)Source codeGastropod shellRootBackdoor (computing)BlogLecture/Conference
Virtual realitySmartphoneFreewareVirtual machineSoftwareOnline helpLecture/Conference
EmulationOnline helpElectronic mailing listSource codeSoftware developerOpen sourceComputing platformComputer configurationView (database)Point (geometry)Patch (Unix)Meeting/Interview
BitComputing platformOpen sourceLecture/ConferenceMeeting/Interview
EmulationFormal languageComputer configurationProjective planeVapor barrierPoint (geometry)Software bugInstance (computer science)Software developerOpen sourceMassLecture/Conference
Distribution (mathematics)Formal languageOpen sourceRight angleSoftware developerVapor barrierComputing platformOpen setLecture/ConferenceMeeting/Interview
Virtual realityOpen sourceWritingSoftwareProjective planeOpen setFormal languageSoftware frameworkTranslation (relic)Vapor barrierEndliche ModelltheorieRow (database)Software developerDrop (liquid)Lecture/ConferenceMeeting/Interview
EmulationVirtual realityComputer hardwareComputing platformTranslation (relic)MassVolume (thermodynamics)SoftwareFamilyOpen setLecture/Conference
Axiom of choiceComputing platformLevel (video gaming)Dependent and independent variablesSystem callReal numberCausalityLecture/Conference
Computing platformQuicksortMeeting/InterviewLecture/Conference
Multiplication signQuicksortVideoconferencingExtension (kinesiology)Meeting/Interview
Computer hardwareBitOpen setGoodness of fitComputing platformStandard deviationDistribution (mathematics)Lecture/ConferenceComputer animation
Entire functionQuicksortInterrupt <Informatik>Different (Kate Ryan album)Keyboard shortcutPoint (geometry)Distribution (mathematics)Independence (probability theory)Standard deviationDatabaseImage resolutionRight angleElectronic visual displayInterface (computing)BuildingLecture/ConferenceMeeting/Interview
Right angleArmNeuroinformatikMultiplicationComputer hardwareDifferent (Kate Ryan album)Computing platformType theoryIntelLecture/ConferenceMeeting/Interview
Extension (kinesiology)Kernel (computing)Point (geometry)Computing platformAddress spaceLecture/Conference
SoftwareSimilarity (geometry)Expected valueSource codeLecture/Conference
Operating systemRight angleLecture/ConferenceMeeting/Interview
Transcript: English(auto-generated)
Hi everyone, so I want this session to be quite interactive. I'm going to try and get you folks to do a bit of work here so the door's there if you want to escape quick. So my name's Andrew, two things you need to know about me if you weren't in the talk I did earlier on. How to contact me to send me
questions, flames, bug fixes, whatever. Second thing is I'm open source manager for the Limo Foundation. Hands up, who knows what Limo Foundation is? Okay, all
right, so Limo Foundation is an organization that's creating a middleware open source Linux platform for mobile phones and it's been going since 2007. It's got very diverse membership across the entire ecosystem of across the entire mobile ecosystem including chipset manufacturers like Texas
Instruments, handset manufacturers like NEC, Panasonic, Samsung, LG, software companies like Opera, Mozilla, network operators like Vodafone, Orange and so
however everything I say in this room entirely my own ideas my own thing nothing to do with Limo. So hands up, how many of you have got a Linux based mobile phone? Okay, yeah so raise your hand if you had a Linux
based mobile phone this time last year? Year before that? Okay, before that? Hardcore. No? Okay. All right, so Limo Foundation's actually been around since
2007. Some of our members have actually got experience of building Linux based phones going back to 2002 or so, but of course it's not the only Linux based mobile distro out there. There are a few others but it's still a
relatively sparse landscape compared to the desktop Linux world where you've got seemingly hundreds of distributions. This is pretty much what you've got in terms of mobile or mobile based distributions at the moment. Don't know if I've missed anybody, anyone feel free to shout out if I have missed anything
obvious. But not everyone seems to be building mobile distributions and it's a bit strange why there's so few mobile distributions out there. Perhaps it's it's hardware based. Normally when you build a distro it's just kind of like building Lego. You can take bits from here and there and put it
together, but in actual fact if you look at the Limo platform we've kind of taken things from here and there. If you look at GNOME Mobile we've borrowed quite heavily from that. Some of the challenges that we face though in building a Linux mobile distribution. Where to find the latest release of the
open source code that we want to use and in some cases where to find an older release of the open source code that we use. When you're building mobile phone platforms you're typically building them over, typically building mobile phone over the course of a couple of years. So if you think back to
2007-2008 that's the code that is probably going to be in a lot of mobile phones today. Finding the source code is is a challenge and finding the license that covers the source code is is a very big challenge and actually making sure that you're compliant with the licenses is a bit
of a problem from the mobile perspective. Another challenge is reference hardware from both a platform developer point of view but also from an open source developer point of view. I had a question in an earlier talk which was basically you know where is the open hardware to go
with the open operating system. No one really apart from OpenMoco have produced open hardware so you're not likely to see an open piece of hardware from perhaps Samsung or NEC or Panasonic anytime soon. So how can you build a software platform if you don't have the hardware to build on? And how can
you make the phone hardware if you don't have the software? It's a real chicken-and-egg situation. One solution might be to look at things like BeagleBoard. Although one of the problems that I found is they can actually be quite tricky to track down so for several weeks when I was trying to get hold of one the websites where I was trying to order them
from were continually saying they're backordered and you'll have to wait. And in perfect timing for this session my BeagleBoard shipped me today which is not a big help. If you had reference hardware how do you actually get it to the right developers? How do we say okay we know the open source
community are going to build a mobile distribution. Who in the community should actually have the reference hardware? How can we identify them? Any suggestions? I'm happy to hear them. Package management. When you're building a mobile platform what should you do in terms of package management? Should you use you know the Debian packaging, the RPM, the
RGZs of all of your platform? So we've seen in Umo we've seen pretty much every possible packaging mechanism being used that you could
imagine. So we've actually seen a platform delivered with all of the packaging mechanisms simultaneously in use which is a bit bizarre. So I guess what I want to get you guys to talk about or to talk to me about is what we
should be doing in terms of a mobile distribution. Should there be a mobile distribution or should it just be another aspect of Debian or Fedora for example? Do you think there should be a separate mobile distribution?
Did someone say no? Why not? So a lot of the things that distribution does is policy based and for a mobile phone the
policy is going to be different for just about every handset that gets created. So it's useful to have a base distribution perhaps but it's a lot more useful to have a system where it's easier to create a distribution. So the
distribution build tools like OpenEmbedded whereby you can define a distribution. You've got base configuration, a base policy and then people building these mobile phones can just build on top of that. That's my opinion but having a Red Hat based distribution or a Debian based distribution is perhaps overkill for a mobile phone. Okay, overkill in
in what way, in what sense? Because a mobile phone doesn't need all of those packages or it's... Yeah, the use cases are very specific depending
on the vendor and even the network operator. So the policy varies so much that it's more useful just to be able to create a distribution easily than it is to have a distribution to strip down from. So it's like building blocks starting up rather than stripping down or that like
so, my opinion. Okay, all right. Does anyone agree with that? Disagree with that? Well, go for it Dave. Oh, sorry. Yeah, yeah, true, yeah. One of the problems
with focusing on mobile devices is that you can't just, it's a lot harder to sort of build, you can't build on the device, you've got to build on your desktop and then port stuff across. So, well, you could develop entirely, well you could develop entirely on the N900 for
example but compilation might take a while, I guess. Yeah, I was thinking that would be quite handy today actually. So, yeah. Dave. One of the
issues I think that mobile distributions are going to have is that specifically limo is that a significant chunk of the platform is
not freely redistributable by the users and until you address that issue the prospect of a successful mobile community distribution is limited, I would say. What do you think about it? How are you thinking about that issue of trying to
make something that your users can redistribute? Yeah, okay, thanks for that question. So, you're quite right that a semi-closed platform is
harder to redistribute. I guess the question there is, do we need to redistribute the platform to the users or just the people developing the devices? But there are certainly huge chunks of, well, 70% plus of the limo
platform is open source, so if we're just working with those open source components and forgetting about the closed or proprietary components for the time being, what would we need to do to to make that happen? Yeah, yeah,
yeah. I mean I don't disagree, there are enormous benefits to having an entirely open platform for sure, but this is more rather than looking at limo specifically in a mobile platform in general, what are the sorts of things that we should be should be looking at? So there's discussion of sort of a policy based
approach, whether it should be stripped down or not. So do you want to just hang on for the microphone? I'm not a core developer on SHR distribution that is
working for OpenMoco, but I participate a bit and I wrote an identical application for it. So basically what I would say there is that how can that be a question? Of course you want it to be redistributable for
everyone, because the inventions will not come from a centralized or from the vendors, but from every free software developer all over the world that can actually build an application. So basically to me that is not a question, it's a matter of fact and everything otherwise will just result in IPAs.
Yeah, yeah, I actually completely agree and this is where the I'm not speaking for my employer bit kicks in. So as an open source person I totally agree if you want a platform to be widely adopted it has to be open and if you want to benefit from you know innovation and bug fixes and so on it
has to be open. So okay, so openness is a given requirement for a mobile platform. So moving on from that, are there other things that we should be seeing in a mobile platform? Sorry I'm losing my voice. Anyone? So should the
mobile platform be hosted by an existing distribution do you think? Should it be a you know Debian or Fedora or or should it be an entirely
separate effort? No opinions? Okay yeah actually that's a fair question why are there so few mobile distributions? Yeah, could be I guess. Do you think there
will be more mobile distributions now that hardware is caught up with or is
you know sufficiently powerful to run Linux? No? Okay. So there any other things that you would really like to see in a an open mobile platform? Yeah.
Before I speak I have a small request can the microphones be turned up a bit? Because there's a consensus on IRC that we can't hear a thing. Is it this this microphone or? Yeah both the microphones. It's on maximum on here.
Okay. As far as open platforms go for Linux mobile distributions, one company I've found has been doing it well recently is PAM. I've saw a few
people here that have the PAM pre and webOS devices. I have one myself in my pocket as well yeah. I think they are doing a good job in trying to spearhead the effort of a completely open mobile platform because they're well compared to other other companies who are being a bit more closed about it.
I remember reading recently about Android for example that wasn't completely open because of certain components although I'm not an Android developer. But at least with PAM webOS it's using open standards. Things like OcTheora and HTML5 and so on just to develop the applications. So it's
a lot more open. I think that with companies like PAM that are notable we're probably going to start seeing an increase in the number of Linux mobile distributions that we have. I mean I didn't catch his name there. He said that
there was only about four which was the four that you listed. But I do think that's probably going to start increasing pretty soon. Okay so would you expect PAM to open up their platform as well? Well the platform already is very open. They provide a number of developer tools to just access the raw
Linux platform that's underlying. They don't try and hide behind the API's and SDK's that they provide. They do allow you to just do what you want with it. They basically provide developer tools and say there you go have fun. And we're probably going to start seeing more of that. Well now that's more to
build on top of the platform. That's not so much to actually work on the platform itself is there. So you can't go and download the the PAM webos source right now and compile it yourself and flash it onto a PAM phone. That's true yeah but I'm hoping that it's going to evolve a
bit. And we're going to see like I said other companies try and follow suit. And because PAM are taking the first step and allowing end users to access the operating system that their phone is running on, which is perhaps a bit more
important than developers just accessing it. Because developers been able to do that in all platforms for a long time. For the end user to have that much access to their own operating system it's probably going to start leading the way to a truly open Linux mobile platform. Okay that's a good
comment actually an interesting thing. And do you think the the users are more interested in the platform because of the the use of open standards and the openness? Probably not many users who would consider an open platform to be a
selling point. Obviously developers would consider that to be a selling point. As far as end users go at the moment they don't really care all that much. But in the Mozilla talk earlier on it was being said that we need to start giving users choice. We need to start informing them of their options. And
that's something that can translate over into mobile platforms as well. One of the guys there was shaking his head about PAM opening the code I think. Somebody was laughing at me. What's your view on, well not just PAM but other
mobile companies opening up the source to their platform? Okay I'm a developer and so I was first impressed as I read someone the first BepOS images was
available and someone looked into it and said oh we used OpenEmbedded to build the whole thing. So I was a little bit impressed. But when the story goes on and as you know Harald Welt has made some pressure so we open up the patches
and the sources we already have in OpenEmbedded because of the GPL violations. Okay okay so and so I don't think we will open some more of the stuff. We even don't give us something back. Open the website we only put the
sources and the patches we already have in OpenEmbedded. So are you saying that publishing the source and patches is not enough? They should do
something more? What should they do? I don't know but that is if you can get the root shell on the primary I don't know. Yeah you can. I believe Joy has also put on his blog something about PAM actually having a
back door on a PAM tray or something. I don't recall the exact details but there was something nasty going on there with PAM doing stuff they really should be doing and he blogged about it and it got in the press and then they changed it. I don't recall the details but it was pretty nasty. The free smartphone people are trying to port the free smartphone infrastructure
to the PAM machine. So basically you can use the phone and free software stack completely but it's still on infancy so we need your help. Okay yeah so one of the things that limo has just done is published a list
of all of the open source that that's used in the limo platform and one of the things they're starting to or we will be starting to do is publish patches but is there anything further that could be done to to help the
open source community from that point of view other than opening the rest of the source which is not an option. Dave, hang on. How about helping developers
working for mobile companies to be better open-source citizens to be better citizens and you know open up a little bit more because you know most of them aren't. Okay yeah agreed. One of the challenges for mobile phone companies
for example is that actually they're working on a platform in a completely different language to their native language. I know this is something that
that is a challenge in you know open source software development in general but if we look at Samsung for example their team are all based in Korea they were talking Korean we've actually seen instances where bugs go through a system and at some point get translated so that the development team
can deal with them and then they get translated back into English to report back to to the outside world so I think the the challenge with introducing open source getting companies to be better citizens is is actually probably more than teaching them how to deal with open source communities it's actually a question of a massive language barrier as well so I think in a
lot of open source projects English is a fairly common commonly used language for them that's that's just not an option and this goes for for many of the handset manufacturers but yeah I take your point that they need to get involved wait for the mic sorry on the language issues it's an
interesting question because I guess it's one that's not limited to mobile distributions any suggestions and that's for the room any suggestions in how we can better engage African South American and Asian software developers
anyone any ideas on how we deal with the language barrier in facilitating open source and open platforms sorry the microphone there I think open source
does a lot better with that than the most of the rest of the world we create a framework whereby people can contribute in different languages and they do I point to the project I'm best known for Apache HTTP D in that
we've got a documentation framework that record that encourages translations of periodically we get one then yeah and it drops us completely out of the blue sometimes likewise if you if you have a book you you suddenly hear on the
grapevine that someone has translated it yeah and this is this is happening in the world it's all it's a model not unlike the open source model of software that once you've written something people will work on it yeah
yes the actual development community happens in English which we do and which is still a barrier I guess what I'm trying to say is I thought about this in some depth when I when I got someone asked out of the blue
completely asked me to write would I do it an intro for the Chinese translation of my book no I and I thought so I wrote something about what an honor it was to be translated into such an important new language and probably the biggest community of non English speakers around and hoped we could bridge the gap for them okay yeah so so perhaps openness is the
answer to or one answer to solving the translation troubles is anyone from open moco here so yeah so I was just going to say so one of the other things
though is that although open hardware is often considered a prerequisite for an open platform it doesn't necessarily always work I guess and is it more a
mass consumer adoption will drive an open platform more than the open hardware do you think I think is that there's no software that works except the MIMO one I mean I bought my open moco two years ago we ported
Debian on it it's not really porting we adopted Debian on it and still is quite difficult to use Debian on it as well as a C chair sometime you miss some cause sometimes the phone doesn't respond anymore I never use anything other than an open moco but at least the open moco is available for everyone
so yeah I don't think it's a problem of the platform itself there are too many too many choices actually okay so that's probably the real problem okay okay so um has anyone else got any suggestions on what we can do
to build an open mobile platform should we all work with MIMO or something new
yeah tell all the press that we have this world-changing device sort of thing that works for Apple every time they have something new and to a slightly lesser extent Google's watchable the Nexus one big press yeah okay it's what get the Americans in it is what they're good at okay all
right um I guess if no one else has any other comments we can draw to a close a little bit early there I think that's actually yeah so so yeah I guess basically open hardware is not necessarily the solution but it does
enable certain freedoms an open platform is is a good thing open standards on the open platform is even better and if anyone's got any suggestions on how we can actually facilitate more open mobile distributions then please give me
a shout and tell me what we can do so actually I've got a question for you obviously with if you're trying to reconcile the entire Linux distribution create one standard for all these devices you do need to be concerned about different chips different display resolutions different keyboard
different interrupts all sorts of different ness yeah what have you guys done to say create the database or sorry the device dependent and the device independent layer so that you're building a standard interface through all these different devices right sort of yeah so that's a that's a good point
actually because the whole device pendants or working on multiple different types of hardware has been kind of solved in the desktop world by having you know builds for different platforms and so on I guess the challenge in the mobile world is that you have to build on a desktop and then
cross-compile for the for the hardware that you're you're aiming for inherently that's the difference between Intel based computers and arm right yeah it's a whole yeah different so to a certain extent we're kind of lucky in that that's partly been solved by the Linux kernel so we're the limo platform sits above the kernel but below the UI layer so we're we're
purely middleware at this point so it's you know it it's then up to each of the manufacturers that take the platform to just compile for you know whatever their target devices so the limo platform itself doesn't actually solve
that that issue or address that issue so you basically take limo and you give it to Samsung to any seed of Motorola and they implement it on their devices yeah okay so you replace the software that they've poorly developed yeah with a standard one and then if they if they do anything that's specific to their device the expectation is that they would then
contribute that back so that others using similar chipsets for example could could benefit from that okay does that actually happen and it's early days yet so yeah yeah so theoretically the Palm Pre could be supported by limo yeah palm could you know come could come along to us and
say we'd like to use your operating system and we'd be happy to help them okay all right if there's no other questions I'd just like to say thank you all and yeah