A Grudging Success
Someone from Koolu called me at buttfuck early this morning, as promised (I always sleep in, and he was in Ontario, so I don't hold it against him). He asked me to try testing it on their port of Android specifically for the Freerunner, which they would be releasing images for in a couple of days. I was not able to get Android running despite some very clear instructions posted recently with Sean McNeil's images. But further investigation today revealed that the initial boot of Android takes an unusually long time to boot (something like ten minutes?!), while it finishes setting up and installing. I may simply have not been patient enough. I'll find some time to try again.

In the meantime, I tried SHR, the "Stable Hybrid Release". This distro used to be an ugly-looking Frankenstein mash-up of the 2007 (gtk) and 2008 (Qtopia) software stacks, but I come to find out that it has now totally changed into what amounts to a community showcase distro for the FSO next-gen Python/Dbus middleware layer currently under development at freesmartphone.org. FSO is the distro that had the great debugging tools that told me what the hell was wrong with the thing in the first place. Well, SHR now simply wraps that in a more complete suite of applications, and a cleaner, faster, prettier interface. I was impressed. The fact that I didn't have to give up Mickeyterm and mdbus was even better.


If there's one thing I'm starting to appreciate as much as Python, it's Dbus. Dbus is the desktop interprocess communication framework that standardizes the ways that programs expose and signal their functionality to one another, rather than just to users.

One thing about Mac OS that I have been impressed by for a long, long time is Applescript, and how applications for the Mac quite often expose functionality, not just to users through the GUI, but to developers through the Applescript APIs. Throw in the recording of Applescript macros that follow the user around, and generate code based on what they do, and writing applications that make use of existing applications becomes marvelously discoverable for novice developers. Instead of having to read the manual, just do what you want the computer to do, and the macro recorder will spit out the code and the parameters, which you can then tweak. I have always felt that that kind of discoverability makes a huge difference when it comes to blurring the line between developers and users. That, of course, has its pros and cons. It empowers users, but I can see where it allows inexperienced developers to make a mess of things, as well. I don't think that's an excuse to keep them in the dark, however. I'd see it as more of an opportunity to learn.

Dbus, then, is the first step toward a standard way to discover what your desktop applications can be made to do in a script, rather than just with you clicking away at it. If some kind of macro-recording function can be built to record the Dbus method calls and signals flying over various busses, and we can begin to approach the rich functionality that Applescript offers. I'm looking forward to seeing it. This is why I'm really happy to see the Openmoko software stack being built around it.


And because of these developing tools, I was able to test more thoroughly and discover some good news: the GSM/phone functionality actually seems to mostly work, and the parts that don't work may not be due to faulty hardware after all.

It turns out that whenever I boot up with the SIM in, the phone registers just fine with Fido. I can reliably recieve calls, and I can reliably send and receive SMS text messages. But as soon as I attempt to call out, It not only immediately kicks off that call, but I am unable to re-register with any carrier or send SMS messages. One thing I noticed was that, before placing an outbond, I was always registering with Fido, using Fido's carrier id number. But as soon as I tried to call out, the carrier ID on further registration attempts was the same, but the phone would somehow identify the carrier as "Rogers Wireless", not Fido. Rogers owns Fido, so it doesn't surprise me that they seem to have pretty merged their networks together. Fine with me, but apprently, it confuses the hell out of my new phone.

The Wifi is still patchy, but it works sometimes. My roommate's PC Wifi connects flawlessly through it, so I may have to live with that. But If I can get service through another GSM carrier that works that last mile with outgoing calls, I think I can appreciate and use the Freerunner just fine.

We'll see.


Speaking of Wifi, I stumbled on some more information about the Kolofonium hack for La Fonera. It turns out, that, yes, the cat-and-mouse game did eventually go to the FON corporate parents. Now, with my (much later - 7.2r3) firmware version, the holes exploited by Kolofonium have been patched up. The only way to own the equipment now is to //build// a physical serial cable to hook up to the thing and flash it.