About one week ago, I received my
OpenMoko Freerunner. This is an openly developed mobile phone that runs purely on Free Software. So this is what I have to tell about it.The hardwareIt was smaller than I thought, and is quite light. My girlfriend says it s ugly, but I m fine with the look of it. Besides being a GSM-phone, it comes with some nice gimmics: GPS, accelerometer, WLAN. The touchscreen works fine, although I don t have anything to compare it with.The softwareThe system it comes with, even after upgrading, is still very rough. It mostly works for doing phone calls and SMSs, but there are a number of unsolved quirks that prevent me from using the Freerunner as my sole phone for now. The suspend mode is left too often, resulting in a battery life of about eight hours, and there are issues with the audio for the conversation partners, who will hear static and echoes. But, as this is free software, there is hope that this will be fixed eventually.DevelopmentThe OpenMoko distribution is based on
Openembedded, which uses
bitbake for building software. So if I got it right, and this is not sure, because documentation is rare and spread, there is the git repository at
git.openmoko.org, which is a copy of the openembedded git repository. This contains bitbake recipies for all
the packages, which includes where they can be downloaded, the package metadata (such as dependencies and version numbers) and sometimes patches. These recipies reference upstream tarballs or subversion URLs. For the native OpenMoko applications, the source is in the
OpenMoko subversion repository.One of the suggested ways of compiling software for the FreeRunner is by using a
toolchain tarball, that can easily be extracted somewhere and used to build the software from the subversion repository, or other (hopefully autoconf ed) software. This builds the binaries, but does not produce proper .ipk files, so no version number or dependencies.The other way is the full openembedded setup, made easy using the
MokoMakefile. This, automatically, fetches and builds everything needed for the cross compiliation and all available packages, producing the same output as can be found on the openmoko servers. Setting this up requires about 6 gigabytes of storage and takes over a day the first time, but then hacking the phone is relatively painless, as it resolves dependencies and is self-contained.The communityFor a free software project, the state of the community is very important. The OpenMoko seems to suffer from a rush of interested people on the mailing lists, so it s hard to follow real development in a mass of frequently asked questions and nice ideas from people who have neither an OpenMoko phone nor wil do any coding.On the other hand, it s not easy for new contributers. I have written some code that make sure the phone can handle numbers such as 0172/123 456 instead of the official +49172123456 in the phonebook and the SMS app, something that other users have
complained about
as well. But no one could tell me where and how I should submit my patches, and the
mail to the mailing list with the patches and the
bug report is unanswered. It is not clear, at least to me, who is responsible for what part of the project quite different to what I m used to from Debian, where there is a clear list of maintainers for each package, and a well known way of submitting patches (by going through bugs.debian.org).For interested users, I have published my branch of the git repository at
git.nomeata.de, and I will hopefully add more features and bugfixes later at least when I find out how to properly contribute to OpenMoko.
It seems to me like basic business common sense not to attack your software’s users, because they’re people who might fund your future development work, when they want some new feature adding or some adaptation making. We should hope that this audience will become co-promoters, co-developers or co-producers of free and open source software in the future.
Indeed, one of the basic lessons at our webmaster cooperative is to try to direct enquiries about software we’ve co-produced to somewhere useful, even if the question isn’t really much to do with us (because it’s about some part we didn’t produce or it’s something we’re no longer funded to work on), or even if we suspect the questioner has done something very silly. This approach has worked well so far. So, just don’t attack your users. It’s not on. Don’t do it.
This basic lesson of not attacking users doesn’t seem to have been learned in software industries in general. Most software comes with a very restrictive “End User License Agreement” that seems to treat the user as if they are a potential criminal and grabs as many rights as possible for the software producer. I expect that Microsoft’s new post-Windows operating system will continue this approach and that the user-tracking possibilities of internet-centred operation will be exploited as far as legally possible. Hopefully, we’ll continue to see pre-installed Linux growing strongly in UK as Bristol Wireless describe.
However, this pales into insignificance compared to many media companies. One of the worst offenders is Bristol-based TV Licensing who send us a “pay up or else, you potential criminal”-style letter each year. If there’s ever a reform that makes BBC subscription voluntary, they will probably reap the seeds of hatred that TVL have sown by attacking their audience.
I have quite a lot of admiration for people like Martyn Drake who are fighting a noble fight in things like TV Licensing - the war begins! but I wonder whether media companies will learn not to attack their users before they cease to exist.
OpenSync update
To give some update on the state of 
One of our most popular products at 
Testing 2.6.27-rc2 with the current released (not development) BIOS on the Foxconn G33M reveals the following:
As 
I left event around 4 PM after Tea break.
As always, we’re doing lots of interesting and significant things and have more good ideas than we have people to do the work. If you have an Eee PC or are planning to get one, or are otherwise interested in our work, some of which has uses beyond just the Eee, here are some highlights:
Earliest Eee models supported in Lenny
Lenny will release with the atl2 ethernet driver and the non-free madwifi-source now works with the earliest Eee models as well, so our patched version is no longer needed. This means Lenny will work with all of the earliest models of the Eee PC: 701 (2G and 4G surf, 4G, 8G) and 900! All we need now for full support in Lenny is to replace the non-free wireless driver with the free ath5k driver when it is ready.
Free drivers
Nick Kossifidis has 
Today I had the honour to prepare and 
I'm in Portland, Oregon for the week where I'll be at 
Just in time, at the end of my 3 week vacation, I released the first public version of smuxi!
For people who don't know smuxi yet: I have been developing over the last 3 years an IRC client that fits the needs I could not satisfy with irssi. I was using irssi over 4 years and it's the best the IRC client I used so far. Most annoying was that it didn't integrate at all into my desktop experience (such as mouse / clipboard / theming / highlight handling). Thus irssi inspired lots of features and commands seen in smuxi.
So here it comes, a flexible cross-platform open-source IRC client for the GNOME desktop, in action:
