Damn, I missed

31 January 2008

I missed the RepRap talk today that I really wanted to see. I saw the Open Source Animation talk by Elizabeth Garbee, but only afterwards did I realise that I’d missed the RepRap talk.

Ah well, the animation talk was okay. It was interesting seeing the demos of software that Elizabeth used to create animations from stills — I had no idea that such software even existed in the open source world. I guess I shouldn’t have underestimated the community. ;)

Wiring the virtual to the physical

30 January 2008

Just saw a great talk by Jonathan Oxer, titled Hardware / Software Hacking: Joining Second Life to the Real World. Jonathan went along the lines of last year’s talk, where he did some wiring of switches to his parallel port, and read the values with scripts written in languages such as Perl, C, and even PHP.

This year, he has wired up those devices to Second Life, and allowed virtual people to interact with those physical devices.

Awesome talk, and if you missed it, it’s highly recommended you see the video (part A, part B — Ogg Theora).

Conduit revisited

30 January 2008

At linux.conf.au 2008, in the GNOME.conf.au miniconf yesterday, John Stowers again presented Conduit, a synchronisation program for GNOME.

Since last year, Conduit has greatly improved — the user interface is much clearer, supports more web services, and even networking synchronisation via Avahi.

Conduit is now at version 0.3.6; last year, 0.2 was “stable”, and 0.3 was in the works.

Firefox 3 SSL error

26 January 2008

Corey, I saw your post about the Firefox 3 SSL error:

Well, I have tried Firefox 3 and I really like a lot of the things that I saw. The “awesome” bar really isn’t that awesome for an Epiphany user, but hey, it is a first cut. The GTK integration really makes me happy. Mozilla has been working on Linux support. Then I hit this dialogue:

Firefox 3 SSL error when accessing a site with a self-signed certificate. Contains no easy way to force Firefox to load the site.

Now I am very angry. Not only did Firefox prevent me from going to site I know is safe, there is no easy to way to say “I trust this page”. And yes, that defeats the point of this dialogue, but the reality for the Web consumer is that I have no control over these kind of websites. Now what do I do?

Corey, I actually support this redesign. Part of the reason the web is so broken at the moment (poor adoption standards, lack of alternative browser support, crappy accessibility) is because web browsers have historically been tolerant of bad coding. In the same way, the main reason viruses plague the Windows world because people are trained in bad practices like Googling for a program, downloading a .exe, and then running it.

Clicking through self-signed SSL warning dialogs is much the same as the aforementioned situations. Make self-signed certificates not seem so insecure, and being able to just clicking through dialogs mindlessly in order to get to the site, is bad from a user trust point of view. The user is being trained to just trust every certificate that they come across, regardless of how reliable the certification authority is.

If you make users aware that these sort of certificates are not good at validating identity, and make it difficult for users to force open these sites (in the same way, I believe browsers should refuse to display pages with invalid HTML in the case of the broken web), then website owners will actually bother shelling out a few dollars for a proper certificate from a reputable certification authority.

Or, even better, you can lobby the Mozilla Foundation for the inclusion of the CAcert root certificate in Mozilla browsers. CAcert is an organisation that offers free signed SSL certificates.

Unwired

25 January 2008

A couple of weeks ago, I bought a Bluetooth headset: the BlueAnt X5.

I’ve been very happy with it — audio quality is clear, and the headset has very good battery life. Also, the headset has a detachable microphone, which allows me once again to use Skype and other VoIP stuff. (I’ve been trying to compile Mumble with not much success. I suspect they’re not properly supporting the amd64 architecture.)

The headset comes with an “audio streamer”, which is basically a small device that can stream audio from a cable to the Bluetooth headset. This means you can use the X5 headset with not just Bluetooth-enabled computers, but anything that has an audio cable coming out of it. The advantage of this is that you can use it with a computer without having to fiddle with configuration on the operating system. You just plug your speaker cable into the streamer. I haven’t tried it, but I suspect the audio streamer will be happy to stream audio to other headsets, not just the X5.

Today, Dad brought home a no-name USB Bluetooth dongle (lsusb reports it as 0e5e:6622) that works beautifully out-of-the-box in Ubuntu Hardy. I’ve been working on getting the headset communicating directly to the computer without the use of the audio streamer. I installed the bluetooth-alsa package, which contains some useful documentation at /usr/share/doc/bluetooth-alsa. It’s highly recommended you have a read of the documentation if you want to get A2DP working in Ubuntu. It basically involves a bit of copy-and-pasting into config files.

For getting PulseAudio working with A2DP, first you will want to get the a2dpd daemon running. Depending on how you’ve set up your ~/.asoundrc, you should have an ALSA device called a2dpd. To get PulseAudio to see this device, you’ll need to enter this into the terminal:

$ pactl load-module module-alsa-sink sink_name=a2dpd device=a2dpd

I tried to get this working automatically by putting it in /etc/pulse/default.pa, but PulseAudio segfaulted on startup. I tried asking on #pulseaudio, but they’re always asleep whenever I ask. That’s the trouble with living in Australia. My mistake. You have to make sure you load module-hal-detect before module-alsa-sink. Works beautifully now.

If anyone has any tips for Bluetooth in PulseAudio, I’d be glad to hear it.