OSiM 2008, prime impressioni

Articolo non tradotto, troppo lungo…

So I have been the last two days at OSiM (Open Source in Mobile), I have heard of many issues, promises, problems, stuff. What do I bring home? Random thoughts, and a heck lot of confusion.

The very proposition “Open Source in Mobile” seems misleading. Open Source is the replacement, industry-buzzword version, of Free Software. Free Software is about licenses which grant people the four Freedoms, it's about copyright licenses. At OSiM I have heard a lot of technology and solutions, very little about software and their licenses.

It is my firm belief that being “open source”, i.e., releasing or using stuff released under an OSI-compliant license does not per se mean a thing. The choice of using Free Software is both a strategic and tactical (technical) decision, but as such is just an enabler, not a problem solver. Am I saying that people in the mobile sector should not strive to release or use software under a Free Software license? Not quite. Simply, there is no magic in the business, hard work is necessary, openness is mandatory, a good understanding of how things work is paramount, especially because there is no past experience upon which we can grow.

A little talk on Limo

Let's take an example discussed in Berlin: Limo (LInux MObile). Limo is a cooperative effort of a number of telephone makers, platform developers and the like, to build a consistent, solid, ubiquitous platform for mobile devices, based on the Linux Kernel. Is Limo Free Software? It is not!

Ok, now I am confused. Isn't it Linux, isn't Linux GPL, the mother of all Free Software licenses? Yes, indeed! And still Limo as such is not Free Software? Yes. Ok, confusion grows. So why are these guys presenting Limo on a conference that names itself after “open source”? Because it is Linux-based. So what? Linux is just the kernel of an operating system, if the things surrounding it are proprietary, it just means that a piece of the whole system is Free, the rest is not. Should we be content with it? Surely not.

So I have asked David “Lefty” Schlesinger, software guru at Access and a very knowledgeable guy with the Limo Foundation, where Free Software lies and where proprietary software lies. He gave me a very good explanation, which I am unable to sensibly report, describing how Limo is made. The bottomline is that most of what they have done is proprietary, although the API are public and open, so that anybody can use them to write applications. Big deal, Windows is not any worse on this front, one would say, we have a lot of Free Software running on Windows' API. His point is, however, that there is no way out, those big guys would not contribute so effectively without keeping a big part of the platform proprietary. I hope they will change their mind.

So the Limo guys have used a piece of Free Software (Linux, some infrastructural stuff, like Gstreamer and other libraries) and used it to build a proprietary platform. Ok, now I understand better what I heard earlier. The person presenting Limo in a keynote speech, Morgan Gillis, said that they have been able to put together half a zillion of patents and to make them available to all the members of the Foundation royalty free. So I waved my hand and asked: “are you able to pass the patent covenant downstream to software recipients”, and the guy said “no”. I have then asked “but how this complies with copyleft licenses and those which have patent termination provisions”, and the guy said with without even blinking “yes, we checked, it is compliant”. Laughter followed when I said “ok, we'll see”, but perhaps it was undeserved humor, as it turns out he was (possibly) right. The thing is all but copyleft, is all but Free Software. Anyway the question about compliance remains open.

Good luck to the Limo guys, good luck to Lefty, they have made a legitimate choice, they have the right to follow it up. Only, this is NOT open source, this is NOT Free Software, it is using Free Software to make other things. Good enough.


The word of the moment is “fragmentation”. Everything is fragmented. People complains that fragmentation multiplies at all levels, it's a nightmare. Fragmentation means that anyone who writes software for a variety of telephones has to deal with a great variance of platforms at many levels. Different kernels, different libraries, different Java, different web browser supporting different Javascript language, and so on. Therefore for each and any implementation there must be a lot of work to fix any idiosyncrasies of the single telephone, with some failures in the process, wasted resources, lead time.

True, but again this has nothing to do with the software being “Open Source”. The proprietary sector is probably less fragmented because it is made basically of vertical applications (one application, one platform or a family of platforms). Better, it is fragmented at another level: all the same, you cannot write applications once and run (almost) everywhere. Fragmentation is here synonym of Freedom and of competition. Only one person has not said profanities against fragmentation: the leader of the Openmoko initiative, Sean Moss-Pultz. Openmoko is as Free as one telephone set can be: the telephone comes for a price, but all software, CAD files of the thing, schematics of the chipset, are Free (capital F).

The solutions to fragmentation, as I see it, are basically two (alternative). As a presenter (Christy Wyatt, Motorola) said, one is having only one hardware platform and only one operating system, dominating all the rest (like in the PC sector). The other is to standardize as much as possible, and of course I mean via “Open Standards”.

So let's put a stop to fragmentation, sit around a table and start de-fragmenting, at least to a decent point. But just don't blame the distribution system, blame the choice many have made (including that of not supporting standards to the fullest length).

Other interesting presentations

We have learned of the efforts of Nokia. Ari Jaaksi has been direct as usual. Mr. Jaaksi is Vice President of Nokia leading the Maemo effort (and one of the Authorities when it comes to Free Software in mobile). According to Mr. Jaaksi, Maemo is going to be a mainstream product for Nokia, not just a technology showcase. Many wonder how this will combine with Symbian and Trolltech news. Both companies have been acquired by Nokia, and oddly enough they are both involved in Free Software production. Symbian has recently announced its operating system, the foundation of many Nokia telephones, will go under a Free Software license. Symbian is a good competitor of Linux in mobile (Maemo is based on GNU/Linux), so let's see what comes out of the competition (fragmentation?).

The other part is Trolltech. Trolltech is the manufacturer of the QT graphic libraries. Maemo uses the GTK, another Free Software library set. Is Maemo switching to QT? There are indications, but per se this is not necessarily a bad news.

At least it seems that in Nokia they are taking Free Software quite seriously, and it is going to stay there, as a long term commitment.

Jay Sullivan from the Mozilla Foundation presented an alternative to fragmentation from his point of view. He himself says that he only has a hammer, so he tends to see any problems as a nail. No surprise then that the solution is to program for a common platform, make the underlying stack irrelevant. The platform is going to be the Web browser. As soon as some eyebrows have been raised in the attendance, the guy explained the efforts to bring the web applications to an acceptable speed (through improving by an order of magnitude — or more — the Javascript speed), and to have a set of API invokable directly by the browser (let's say, something like firing up the camera and capturing a video stream).

As an Open Standards advocate my only fear is that as soon as there is a common base for web applications in the mobile, somebody will try and play the free runner. With the existing standards we experience many browser and platform to be plainly incompatible for no apparent reasons. Some would also say that standardization is stifling innovation by providing too many constraints. Nonsense, if you ask, because a non-standardized environment burns so many resources just to test any environments and to hack for a particular platforms. Mozilla's Firefox (and Opera, and others) testify how one could innovate and remain standard-compliant.

Intel has presented another platform: moblin.org, and gave me a nice jersey (thanks). Moblin is an adaptation of GNU/Linux to the mobile devices, making good use of the Atom CPU. The plus of the platform (entirely Free Software) is that it is blazingly fast: fire it, wait five second, bingo, you have the GUI up and runnig. The downside is that it is only an Intel (x86) platform, but it can be ported (and it seems there are ports) to ARM and other CPU's.

The hardly unnamed absent

Mobile phones, innovation, platforms. Wait a minute, why has anybody avoided mentioning Apple's iPhone? My fault, I wanted to save it for last. Apple iPhone, like it or not, has raised the bar of what a mobile thing must do to compete. It is awfully expensive, deficient in many parts one should expect in a mobile phone, it is painfully slow at times, some other times it simply does not work. But it is in almost anybody's dreams!

And it is not Free Software. On the contrary, it is as closed as it could be. Apple controls all the hardware and software stack. Apple controls the Apple store and what a user can or cannot install. All software makers which sell applications on Apple Store must sign an NDA and accept quite strict contractual rules. Finally there are reports of Apple having installed an application killer (not only a killer application!) which has the alleged scope of eliminating any non-approved application.


In a way, though, Apple has raised the bar for all competition. Only, guys, be a little more relaxed, come on!

Mobile me, mobile we

One of Apple's killer features for iPhone is syncing the PIM information on the phone, by means of a facility called “Mobile me“. Again, all proprietary, using non standard protocols, working only on iPhone and with few services (Apples, Microsoft Exchange?)

This brings me to a competing solution, which happens to be fully Free Software (albeit under a dual license regime), which works with almost any smartphone out there. Funambol's CEO, Fabrizio Capobianco, has chaired a parallel session and given an interesting speech on how the fact that Funambol has a community development is important in bringing solutions to the market. All of this enabled by the license: Funambol uses the Affero GPL.

To me (and not only to me, of course), Funambol is a very interesting initiative that must be watched quite closely.

The hardly nominated absent

The most noticeable missing guys were Microsoft. They came along last time I have been at OSiM, which was two years ago, where I met Mark Lange, a very fine lawyer, a man you'd love to disagree with. I wonder why they were absent this time. Of course they are not doing any Free Software also in this sector (they have the “Shared Source” approach, but that hardly counts), nonetheless it would have been very interesting hearing their opinion.

After all, you still cannot solve any equation in the IT market leaving Microsoft out of it. And as long as the mobile part has to rely on and sync with software in the enterprise IT, this means a lot of Microsoft products, like Exchange and Windows.

See you next year in Amsterdam!

Tipo di Entry: 

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.