Month: April 2008

Recording with Palm TX

Posted by – 30/04/2008

I gave Brenda a Palm TX, so she can better organize her life. As a plus, it is wi-fi aware, so all sorts of interesting things can be done. Sadly, though, it doesn’t ship with a microphone (I still wonder why… it’s large enough to hold an embedded mic, just as Palm smartphones have). Lately, Brenda was talking about recording some sessions or holding her thoughts in one of those tiny memo-recorders, and this promptly stroke me as “another gadget to stuff her purse”… then I remembered having read that Palm TX have all that is required to hold a microphone, but the microphone…

I retrieved the Palm Multi-Connector pinout from Wikipedia and realized that all I had to do was attach an electret microphone to 16 and 17 pins and there we go. I ordered a charging cable just for the connector and here are the pictures I took of the building process:

The tools and pieces I used:

It’s easy to disassemble the connector. Take a close look at the attaching plastic hooks… there are four of them: two outside, near the pins, and two inside near the cable. Last two are harder to detach, but once you’ve detached the first ones, just hold the two pieces of the connector and easily shear them: one side will detach first.

The microphone have two terminals that already hold enough soldering tin. Beware to solder the “blue” cable (the one that goes with pin 16 – AGND) to the marked terminal of the microphone, and the “red” one (the one that goes with pin 17 – MIC_IN) to the unmarked one. (In the picture, my cables are all black, but the wires inside are “red” or “blue”. I could not get a good picture of the terminal markings).

I have dissolded the charging cables from pins 1, 3, 5, and 8, but that is up to you (you can just cut the main cable and leave the pins untouched). Also, while soldering the microphone cables to the connector, beware of using as little tin as possible. It’ll be even easier if you remove the pin from the connector… it’s easy to do if you have small pliers and careful fingers. Try first with one of the other pins that will not be used. Solder the “red” one to pin 17 and the “blue” one to pin 16.

I cut a piece of an old earphone cushion to close the charging cable hole (and to protect the microphone). Here is the final result:

Now, for the recording software, there’s a nice free software that does that perfectly for PalmOS: Better Recorder. Now Brenda can record her mind!

Places I’ve been

Posted by – 30/04/2008

I liked what Marlon have done listing the airports he has been at, and I decided I should list the places I’ve been at. For now I listed some I could remember, but surely the list will grow as I can put my memory together. You can check what I have so far here:


View Larger Map

I’ll never expire comments on this article… if you met me someplace it’s not listed, please leave a comment and I’ll update the map.

Considering git…

Posted by – 23/04/2008

It has been some time since I’ve blogged about Subversion and how I was finding it useful. I’ve been using subversion and svk since about that time and I love it. Svk is really great and although it is more an off-line version control than a distributed one, it goes a long way for a former cvs user like myself. Lately, I’ve been considering git (isn’t everyone?).

It was not until I moved to my own hosting (thus having to maintain two subversion trees: the new one and my company’s old one) that the whole “having a central repository” started bothering me. Merging became complex, since now I have two “central” repositories, and while svk makes it a lot simpler (I just keep a mirror of the two repositories and use svk to merge between them), I began to realize that there should be a better way… I’ve watched (again) Linus’ git lecture at Google Tech Talks and decided I was going to try it.

First thing that called my attention was how easy it is to share a repository: just copy the .git directory somewhere public (an http server, for instance), and there you have it. This is much easier than configuring modules for apache, or configuring svnserve (and I’ve done that a lot in the past 3 years!).

For the subversion-only user (one that has never used svk), the carry-all-the-repository-with-you thing may sound strange, but with svk, I was used to “mirror” the repository, so that’s not an important change to me. I could not compare space utilization (and git is advertised for being more space-efficient) yet, but so far the git repositories seems to be a little smaller.

I already found it easy to branch using svk (in fact, I used to do that a lot), but git branching is really a plus! You can switch back and forth inside the same ‘working copy’ (actually I don’t think it’s correct to call that a working copy) and merging is rocket-fast! Maybe if svk could easily merge between two distinct users without requiring a subversion repository (think about the defunct svl), I would not have been trying git, but…

One thing that is really missing is the partial checkout. Actually I liked very much that svn/svk “feature”, and the clean and natural way which they treat directories and files. I understand git is designed with other priorities in mind, but, right now, that feels like missing to me. Maybe it’s just a svn/svk habit.

All in all, I found git very interesting… I am not ready to switch from svn/svk yet, but I’ll try git in small new projects and see if I can get used to it.

So… it’s over

Posted by – 20/04/2008

And fisl9.0 is over. This was a great one… not only because of its size, but there were so many things at stake that I am positively surprised such a few went wrong. We’ve got a lot to think until next one…

To begin with, there’s a major difference between this “inter-fisl” time: we already know where next’s going to be and how much it’ll cost us. The event-center cost, as you can imagine, is the most expensive part of our budget, and knowing that in advance, I believe, will play an important role in our planning. Also, we got more time to plan for the next one: it’ll be in June, not April, as usual.

Moving back to PUCRS (from FIERGS) as fisl’s event-center – now I can say – was scaring me to death. PUCRS is smaller and we were foreseeing a huge event. Guess what! The event was bigger than first imagined, and PUCRS held it without major drawbacks. Sure, lectures were crowded… but they were already crowded at FIERGS anyway… The expo had more people than FIERGS, but I am sure our sponsors did not think that was a problem 😉

Of course, not everything went OK… I think we had two important problems to solve for next fisl: our network must be born already correctly configured for the streaming, and our wireless infrastructure must be stronger. The first one I already wrote about and I think it had to to with the inexperience of the network people this year: apparently they’ve put everything under a crappy load-balancing and while this might work for web-surfing or sending small amounts of data, it just hurts anything that should be continuously transfered, like real-time videos. Here is a piece of advice: network planning involves the “right” amount of everything (the right amount of cables, the right amount of bandwidth, the right amount of uplinks), including the right amount of configuration! I think our network problems were more a case of overconfiguration than one of lack of skills.

The problem with the wi-fi is deeper. Maybe I don’t completely understand the technology (and people handling it were not very helpful). How do you manage an event like fisl in the wireless front? PUCRS have provided us with enough wi-fi to hold 1500 concurrent connections (so they say), and have reported we never had more than 400 concurrent connections… and yet, as many people have reported, to connect our laptops was a hassle. I don’t have a laptop, but I’ve seem a lot of my friends trying to connect with no luck. During the second day, PUCRS wi-fi technicians told us that the problem was the Access-Points people brought to fisl (according to them there were more than 30) that were, somehow, messing with the channels they were using. So we reserved some channels to PUCRS and asked people to use other ones. I tend to believe that the majority of the people listened to our appeal and reconfigured their Access-Points… but no improvement were reported.

So, either PUCRS cannot hold more than 400 concurrent connections in their wi-fi system, or there’s some other problem we’ve not identified. What is the right thing to do? To forbid people to bring their own Access-Points? No way I am going to see the day fisl organizing committee will disallow people to carry their own communication devices!! There must be a better solution… I think I am going to study wi-fi at least not to be easily fooled around by this “messing the channels” crap (please, comment to this article if you have any ideas).

Finally, people are dreaming of a 10-thousand people fisl next year. I think it’s not more than a dream… but hey! When we made the first fisl, there were only 300 people… it has been a 9-year long dream… Maybe that 10-thousand one will come true, as the previous ones did. I will just add one more problem to our list: the first day long lines to get people’s badges must be addressed in a creative way…

Thanks to all that came to fisl. See you next year!

fisl9.0: 2nd day

Posted by – 19/04/2008

They’ve came down with a participant number for fisl9.0: 7417. That’s it… over SEVEN THOUSAND! I was told it’s still an estimate, since they haven’t merged the databases yet (anyway, that’s the official number so far). This fisl is huge! One more picture of the crowd, just for the record:

Now everything is fine. TVSL went online with no problems (once the network was fixed) and remained broadcasting the event the whole day long. All was so fine that one can think it’s even boring! I spent half the day attending our company booth and the other half trying to help the organizing committee. While walking around, two linux-driven home-made robots called my attention. The first one serves water to the guy in front of it:

The other one just walks around:

On the Debian side of the “trench”, I met João Eriberto Mota Filho, Debian Developer wannabe who asked me to sign his key. I’ve seem he before, since he’s a frequent speaker at fisl and welcome him as a future developer. Meanwhile I was “reminded” by faw that I still owe him my signature… since DebConf4!! I signed so many keys in the KSP we held then, that one or two might have been missed… I intend to fix that RSN 😉

Wow! Fisl exploded!

Posted by – 18/04/2008

I have never seem so many people at fisl. Seriously! All the rooms were crowded, all the time… same as the expo and the user-group area. Our official counter were stopped at 6667 participants, but that is just because we’ve taken the registration database off-line in order to process new registrations on-site. By the end of fisl all the numbers will be summed up, but I heard rumors of 7 thousand people! Awesome!

I spent this first day mostly in the TVSL room. We’ve had problems with the network all day long. PUCRS’ technicians and our link providers’ personnel were trying to figure out where the problem seems to be. Being stuck with a non-functioning stuff, to me this day was mostly a sequence of frustrating attempts to get over the blockage. By the end of the day I was told that they were close to a solution… It reminded me of our second year trying to stream fisl, as was best told by Fluendo’s Thomas … This just leads me to an unavoidable conclusion: our network sucks! I had to hear from a Brazilian friend who lives in Europe how wonderful her Gigabit network is… grrrrrrr

Bored as I was, I went walking around. That, surely, were the best part of fisl so far: meeting old friends. I had the most pleasant chat (yet quick) with Otavio Salvador and Felipe van de Wiel, both friends of long time.

I took some pictures of the crowd:

And, since I was not “available” in the morning, when the huge crowd was arriving, I took a picture of a picture taken by Marlon of this initial moments:

(It is not really clear in this picture, but beyond the red gates there were as many people as inside)

Procura-se talentos em software livre

Posted by – 16/04/2008

Estive no fisl9.0 esta manhã e confirmei algo que já está se tornando rotina em eventos de software livre, especialmente nos últimos fóruns: As empresas em geral estão procurando talentos em software livre como nunca antes ocorrera! Agora eles chegam a anunciar como uma parte importante do seu stand no fisl… Eis uma foto do stand (ainda sendo montado) da globo.com (lê-se “precisamos do seu talento” bem grande, com um fundo de nomes de tecnologias de software livre):

Minha empresa também é patrocinadora, e terá um stand no fisl… e também estamos contratando. Se você entende de infra-estrutura de softwre livre ou de programação, passe lá para tomarmos um café… mesmo se já tiver visitado o stand da Globo 😉

fisl9.0: T-48h

Posted by – 16/04/2008

We are at T-48h of fisl9.0, and everything starts to take form. I spent the whole afternoon getting our icecast infrastructure online. We’ve got 6 servers, all with different bandwidth and not all with the same architecture… The debian boxes were prevalent (4 of them – 1 amd64 and 3 i386), all of them etch. We’ll be able to follow the statistics at tvstats website (not all the servers are enlisted yet).

In this process I found an evil bug in debian’s icecast package, I will have to debug to understand it better: when relaying from a master server (which is our set-up), if the source stops feeding the master, the relay segfaults. The strace shows it just segfaults while pooling… I got no explanation for it and I’ll have to dissect it if I want to build a proper bug report. In the meantime, I installed the sarge-backports version, and it worked… so, the bug report will have to wait until after fisl 🙂

I took some pictures to show how’s everything being put together. Take a look:

And some of the infrastructure crew:

See you all there!

e-fisl

Posted by – 14/04/2008

I’ve been involved in TVSL (Free Software TV – a streaming infrastructure for the fisl) since its beginning… This time, while searching a way to embed IRC into a webpage for a internal project, I discovered mibbit – an interesting IRC widget. Then I thought, “It would be interesting to have that for fisl”… this is how I put together e-fisl.

e-fisl is just a webpage with an embedded cortado player (via the beautiful MetaVid library ) and mibbit widget. I put it together just a few minutes ago, but I suspect it is going to be interesting to watch how people use it (the mount-points for the rooms are still not available – as you’ll notice).

I am planning to have a bot in the #fisl channel to collect questions from the “remote attendees”… I just have to check with the rest of the organizing committee if it will be possible to read these questions for the speaker during the session. It would be great…

Finally, I am back

Posted by – 12/04/2008

I can’t possibly summarize all that has happened to me since my last post. The short explanation is that I had to serve in Brazilian army for a year, and that kind of disorganized my life. While I was in the army, I was forbidden to blog, and when that year ended, my life was too messy to keep up and I’ve been procrastinating my come back! Now, finally, I am back.

Just in time, I would say. fisl9.0 is about to begin.

For this come back, somethings changed: my blog engine (from Rodin to Mephisto), my Gallery (it’s gone… sorry!), my domain (I’ve moved from my company’s webserver to my own), I will be publishing some entries in portuguese (in a separate section)… Hope you all enjoy.

Happy hacking.