OSDL's Mobile Linux Initiative

Rob writes "The Open Source Development Labs has introduced a new initiative to accelerate the adoption of Linux in the mobile market by providing a forum for device manufacturers, network operators, and application developers. Like the OSDL's other working groups, the MLI will provide a forum for creating requirement specifications based on existing implementations and invest in existing and new mobile Linux projects to identify and fill gaps in the open source operating system's functionality."

Comments

Hmmmmmmmm

[Anonymous Coward]

Like the OSDL's other working groups, the MLI will provide a forum for ..slashdotting the hell out of innocent websites...

Desktop Linux Improvements

[CyricZ]

Would any of the improvements enhance the desktop Linux experience? Indeed, applications designed for the minimalistic hardware of many handhelds could potentially perform amazingly on a high-powered, modern desktop.

Re:Desktop Linux Improvements

[Don_Casper]

I think many of the programs already exist for high end desktops they just want to make less resorce intensive programs for handhelds. IT might be useful for installing on a 90 mhz 486 with 16 mb of ram i suppose. You could give a computer running software like that on a computer like that to impovrished people.

Re:Desktop Linux Improvements

[FidelCatsro]

Yes but saying that , you could just use something like fluxbox or even TWM(to go to an extreme) , Which both perform brilliantly on today's hardware .. They even perform wonderfully on 10+ year old hardware .
Even things like Xfce perform wonderfully on old hardware ., So you can imagine how well they perform on up to date hardware .
Performance is not something to worry about in desktop environments, we don't need better performing WM/DEs (even KDE can run fairly well on a p2 ) .

What we perhaps need is an improved UI in some things (personal preference ).No gain in this sector will likely appear from this initiative , considering the way most Mobile devices handle input .

Re:Desktop Linux Improvements

[Tomchu]

Are you kidding? Both KDE and GNOME feel like a sluggish mess compared to Windows 2000/XP on the same hardware. Hell, even OS X on my 1.33 GHz iBook feels faster than either KDE or GNOME on my 2.4 GHz Athlon 64 with 2 GB of RAM.

Re:Desktop Linux Improvements

[FidelCatsro]

I am sitting using KDE on a 400 MHz G3 right now , whilst it is not exactly lightning it certainly is not that painful to use . .It runs perfectly on my 1.25 GHz G4 and on my 1.9 Ghz Athlon laptop with 512MB of ram ..

Re:Desktop Linux Improvements

[chrismcdirty]

KDE is lightning quick as far as a 1.5GHz Celeron goes on my laptop. But Windows XP is slower than molasses. I guess it varies from setup to setup.

Re:Desktop Linux Improvements

[Anonymous Coward]

IT might be useful for installing on a 90 mhz 486 with 16 mb of ram i suppose. You could give a computer running software like that on a computer like that to impovrished people.

Why --- aren't they suffering enough already?

Re:Desktop Linux Improvements

[andyr0ck]

from TFA: "forum for device manufacturers, network operators, and application developers" the other ones i see possibly making a difference. i run a linux laptop and although i'm happy with the level of functionality i've got out of it, i'd rather be able to get every bit of functionality out if it. maybe just the way i read it, i suppose.

What sort of gaps are they trying to fill?

[CyricZ]

The article itself isn't very technical. Can anyone shed some light onto what sort of gaps they're trying to fill?

It does list some of the technical areas to be worked on:
real-time, power management, security, memory footprint, and fast boot functionality

However, what specifically could be improved in each area? Would such improvements be of any benefit to server or desktop applications in a significant way?

Re:What sort of gaps are they trying to fill?

[Samrobb]

You can probably get an inkling of the sort of problems to be tackled by taking a look at the CELF Developer's Wiki. If you enjoy specifications (And my! Who doesn't?!?) you can take a look at the final CELF 1.0 specification. I suspect that a lot (if not all) of the areas that CELF has examined will have to be addressed by OSDL/MLI as well.

Which just raises the question of why OSDL is putting together Yet Another Industry Forum when one already exists. In particular, all of the comapnies listed as initial members in ODSDL/MLI - MontaVista, Motorola, PalmSource, Trolltech, and Wind River - are already CELF members. What do they hope to accomplish via OSDL/MLI that they couldn't address as part of CELF?

What would be nice...

[dslauson]

... is if they could provide a nice IDE and debugger for embedded software development like Wind River does for VxWorks.

Re:What would be nice...

[tabbser]

WindRivers vxWorks debugger is GDB, just like Linux.

Linux's weak point?

[Anonymous Coward]

Linux is huge in the server market, dominates many embedded system applications, and has made great strides on the desktop. But in the handheld area, it seems to be one step forward, two steps back.

First Sharp moved the Zaurus from a proprietary system to Linux... ...but in a few years, the Zaurus was withdrawn from the international market.

Then our hopes were raised with PalmSource's high-profile Linux move. But it seems to be quite delayed. And what's worse, device manufacturers -- even Palm (the corporation formerly known as PalmOne) -- are running away from PalmOS as fast as they can.

Linux is almost the kiss of death for handhelds, or at least the last refuge of a desperate company. Can this initiative overcome that terrible legacy?

Re:Linux's weak point?

[CyricZ]

Perhaps it's more an issue of trying to focus on too many markets at once. Quite opposite to the UNIX credo of doing one thing, and doing it well, it seems that some people want Linux to do everything these days on all sorts of systems. Handhelds have vastly different needs than servers, and desktops differ from both of those. You're going to run into problems when you try to make one piece of software perform excellently on such varied environments, each with vastly different requirements.

Re:Linux's weak point?

[Bastard of Subhumani]

some people want Linux to do everything these days on all sorts of systems

Linux can do that. Not the same linux, of course...

Metrowerks involvement?

[CyricZ]

It is widely known that Motorola, who is involved in this initiative, is linked to Metrowerks (or what was Metrowerks) via Freescale. Metrowerks is well known for developing C/C++ compilers for various platforms, including many embedded systems.

While GCC is a fantastic compiler, the Metrowerks compiler has often been amongst the best optimizing compilers. That said, could the performance of handheld devices be improved by compiling the Linux kernel and most Linux software with the Metrowerks compiler, rather than GCC? Considering the kernel's use of GCC-specific features in its code, it would of course not be an easy task.

Re:Metrowerks involvement?

[m50d]

That said, could the performance of handheld devices be improved by compiling the Linux kernel and most Linux software with the Metrowerks compiler, rather than GCC? Considering the kernel's use of GCC-specific features in its code, it would of course not be an easy task.

Intel managed to get ICC compiling the kernel (though I hear it took them some work), so it's not impossible that Metrowerks could do the same.

Re:Metrowerks involvement?

[CyricZ]

Intel's compiler on Linux has a high degree of compatibility with GCC. Of course, that's necessary in order to be useful on Linux systems where other libraries have been compiled using GCC. Plus many Linux applications are developed using GCC, and hence GCC often is known to compile the code. So considering their current compatibility, it's not surprising that they were able to compile the kernel with their compiler. I'm just not sure how much compability Metrowerks' compiler offers with respect to GCC extensions.

OSDL

[Anonymous Coward]

Every time I ride past the OSDL on MAX, I say a little prayer for Linus and his team.

High time!

[LaughingCoder]

As a (former) mobile software developer, I think this is fantastic. Symbian and Windows Mobile are currently the two mainstay "open" platforms for developers. Java is making inroads, but no two phone's Java VMs are the same, making it a nightmare for developers (write once, debug everywhere). Symbian suffers a bit from this malady as well - each phone model resulted in a some tweaks to the software, and more often than not, a special executable. It should come as no big surprise that the best platform today for mobile developers is from Microsoft. The tools are quite good (and free), and the various instances of the platform (Smartphone, Pocket PCs) are sufficiently similar from an application's persective that only 1 source tree is needed. Further, the same executable will run on every device as long as it is a compatible CPU like the ARM - which is what almost all of them are. Linux in the mobile space would be a great thing - especially if a decent set of tools (and libraries and FOSS applications) were available to accelerate the process.

Re:High time!

[Anonymous Coward]

Uh, Free? Similar?

I don't know, last I remember PPC had numerous compat. issues like the symbian devices.

Also battery power can be a killer.

Re:High time!

[LaughingCoder]

I had a single executable that ran on PPC 2002, Windows Mobile 2003, Smartphone 2002 and Smartphone 2003. If you just take the time to query the provided and documented APIs you can determine what platform you are running on and make the appropriate runtime decisions. Symbian did not provide these query functions, which forced us to make customer EXEs for every model. And yes, all the tools I used to release commmercial products were free downloads from Microsoft.

Using Python and Ruby for handheld development.

[CyricZ]

Will they be looking into the use of Ruby and Python for application development on such handheld devices? Indeed, many cell phones use Java-based programming environments. Considering how much better that Python and Ruby often perform than Java for desktop applications, it would seem plausible that they may have a place on such devices.

Re:Using Python and Ruby for handheld development.

[CrankyFool]

That's already a reality for some hand-held devices -- you can get Python for your Symbian-based phone. In fact, I can (with great pain, due to the keyboard issue) type "print 'hello, world!'" on my Nokia 7610 with the expected results.

Sounds like CELF

[Samrobb]

From the sound of it, it seems like they're going to be covering a lot of the same ground as the Consumer Electronics Linux Forum (see the CELF home page.)

The Consumer Electronics Linux Forum (CELF), a California Non-Profit Corporation, is focused on the advancement of Linux as an open source platform for consumer electronics (CE) devices. The CELF intends to operate completely within the letter and the spirit of the open source community. The CELF is a place to come and discuss various issues that are of particular importance to the CE industry. Through an open process, the CELF members will clarify and codify certain requirements to be addressed in open source software. Thereafter, the CELF will evaluate any open source submissions as to their effectiveness and responsiveness to the requirements. Open source submissions accepted by the CELF Architecture Group and Steering Committee will be incorporated into the CELF source tree, which itself is open to the public.

There's some interesting stuff coming out of CELF, if you're interested in that sort of thing. IIRC, Matt Mackall (the author of the linux tiny set of patches) is now working for CELF. From other CELF members I've talked to, I can say that they seem to be a pretty techically oriented bunch. The individual CELF members aren't marketing types trying to push OSS developers to do their work for them, they're developers who have a real interest in pitching in and helping to make Linux a useful OS for consumer electronics.

The device I want

[pr0nbot]

The device I want is an i/o box for a media server. Feature wishlist:

. <$200 so I can buy several for various rooms
. wifi (maybe with base station capabilities)
. screen (large enough to host a menu system)
. enough grunt to play streamed audio, and possibly video
. audio out
. video out (maybe)
. expandable storage
. as small as possible
. rechargeable battery and mains
. linux/homebrew SDK

Something like a cross between an iPod, a remote control, and an Airport Express with AirTunes.

The idea is that I have my central media server but can use the console as an output and to control it (select what I want to listen to without a separate remote).

The PSP was so close, it fails on the last of the above conditions.

Finally - a good PIM device

[MarkSwanson]

What I'm hoping to see is an open solution that I/everyone can contribute to so we have a good open standards compliant PIM software.
Currently all of the folks that write PIM software for cell phones have created their own flavours of iCalendar and vcard. They improve their implementations with each new cell phone and each revision of each new cell phone, but quite a lot of limitations and problems still exist. It's hard to support moving targets like this and the targets have been moving slowly for over 4 years.

Personally (as the author of a SyncML solution and someone who has to support n phones/revisions across the globe) I haven't seen any cell phone devices use anything but the older deprecated VCARD standard. I also don't remember (very tired though) seeing any devices that support the newer IETF iCalendar standard. I believe an open solution would be able to get out an updated implementation of the newer standards and enable a much higher level of functionality and interoperability. I believe a better PIM foundation will enable more interesting and useful applications and services to be built.

Cheers.

The initiative

[Jozer99]

The initiative will be ignored by every manufacturer and even consumer, as each will want their own pocket distrobution, and pocket linux will be back where it was, nowhere.

Java does not outperform Ruby in the least

[brunes69]

The only reason you percieve it being better on the desktop is because Ruby and Python use bindings to an already native GUI library, while Java has it's own which is not native, so of course it will seem slower.

This is not an obstcale on mobile platforms since MIDP defines standard GUI objects which are implimented in native code.

Java actually runs faster than Ruby or Python in the back-end because it is compiled code, whereas Ruby and Python are interpreted (notwithstanding JIT compilers). On a mobile platform though, there would be no benefit to any of them, except that Java already has a huge developer and application base in the mobile arena, so it would win out.

Re:Java does not outperform Ruby in the least

[Taladar]

I think Ruby or Python would win if speed isn't an issue with any of them because of the pita that is Java the language.

Does OSDL get it? Who Knows?

[viewtouch]

OSDL's mobile initiative will fail if it doesn't require that devices contain an X server than can automatically discover X client applications.

Troll?

[Anonymous Coward]

Troll? Are You Kidding? Get a Fucking Clue, why don't you?

carriers

[Anonymous Coward]

One aspect is to be taken into account : carriers
Carriers want to have a common basis they can port their apps to and/or ask manufacturers to port on cheaply and fast.
Symbian more or less means nokia, so it is not a very good option (nokia owns > 50% of symbian)
Windows CE is even less a good option, as carriers are just data link providers

So let's say carrier X has a nice feature to add, IM for example.
Today for midrange phone it means doing X ports for all the platform/OS available
With linux, they can provide a sample application.
That does not solve the big middleware issue that is telephony API which is not standard at all, TAPI being very uncomplete

My vision on this, only linux and windows CE/pocket pc will survive.

Embedded Linux

[Rac3r5]

I have a couple of questions regarding the embedded linux OS.

what are the min requirements in terms to get embedded linux into a chip?

What microchips would you recommend using?
There seem to be so many embedded linux OS's which one is a decent one? I don't really need an fancy gui output capability, maybe output data to an LCD. What I mainly need it for is just to control I/O on various pins..

Open? Probably not

[gtada]

I'll tell you why I'm not very optimistic about this.

Motorola, one of the members of this new consortium, has a line of smartphones (A780, E680, E680i, etc.) that run a version of Linux from Montavista (also part of this new consortium) and uses Qt. Sounds good so far, right?

I own an A780, so the following is my experience with the device and Motorola. It was a bitch to get the kernel source. Thankfully, I found it at MotorolaFans.com (I think it originally came from a Chinese site). Motorola is keeping their platform very closed, and it's next to impossible to get a native SDK instead of the slow and not-so-featureful J2ME. And to add to that, the location API (for GPS) and other libraries are locked for only Motorola-blessed developers. The fun stuff is locked away. Their developer site, Motocoder.com, is equally useless with no community forums but just Motorola propaganda. Boo.

IF Motorola et al. actually open up the platform then I'm all for it. Otherwise, I don't think this will go anywhere. The "kiss of death" here is actually keeping things so closed that developers can't do anything great with the platform. I could kiss the Motorola management for releasing cool hardware... but I could piss on them for limiting it with shitty software and developer support (to clarify I'd really only piss on them). :|

When it runs on an 02 XDA

[Anonymous Coward]

When linux can run on an 02 XDA and we can still use all the phones wireless/GSM/GPRS/Phone/PDA features with a decent gui and apps to support then linux could be viable on the desktop, but for now there have been very small inroads to making handhelds really viable with linux