Hacking for fun and profit

I’m in the middle of a 2 day long hackathon with breno. I’ll blog about this later, but now I want to talk about an amazing peruvian powerdrink called Vortex. Try it, i’ve been drinking RedBull and coffee and I was falling asleep here. Two slurps of Vortex and I’m a hacking machine again.

It is a mixture of coca leaf, caffeine, taurine and guarana. Dynamite in a bottle.

I’ve been considering some change in my habits. As some people may know, I am anti-bloat, but I’ve been considering using gnome as my desktop, specially after a week installing, configuring and USING gnome boxes at CONEIS. It didn’t feel “bloated” at all (maybe because I configured it), and it wasn’t bad at all. I have installed gnome at elysium, my box in Americatel, so we’ll see how it goes…

I’ve also been considering trying that other editor, mainly just to give it a try, but also because there is some interesting stuff for coders available for emacs. If I like it, I’ll stick to emacs, if I don’t, I’ll port that stuff to my beloved vi (I’m already highly productive with it).

I’ll keep you informed about the results :-)

A new GNU/Linux distribution has been released: CuasiLinux. According to his author, Ricardo Arroyo AKA “El Cuasi”, this will be the distro that will bring balance to the force. Finally a real competitor to debian?

You can see the moment of the presentation of CuasiLinux in CONEIS:

Tomorrow I’ll post more pics of CONEIS, stay tuned :-)

Yesterday, Marcelo Branco closed CONEIS with an excelent talk about Free Software and his experiences in Brazil. It is amazing how everything is going in Brazil, I hope someday we can do the same in Peru. I think this was the best talk of the entire event and the students loved it, EVERYBODY wanted a pic with Marcelo.

After that, we went to the backstage and we had a nice conversation with Marcelo and Enrique Chaparro about free software and its use in the government and in the military, and of course about debian :-). Marcelo was quite impressed by the amount of people in CONEIS, he said that it was the second biggest free software event he has ever seen. Note that CONEIS wasn’t officially a free software event, but we turned it into one ;-)

Then we moved back to the auditorium for the final ceremony and the election of the university that will host the next Coneis. The next Coneis will be in the city of Tarapoto, in the middle of the peruvian jungle. I said to Marcelo that next year we’ll meet again in the jungle, and he said “I hope they call me”. What a great guy :-)

Overall it was a very good congress. It had some disorganization but mainly because the students of San Marcos were completly alone in the organization of this BIG event. I would like to thank Jorge and Katia for their great support, and for working so hard.

Se you at the next Coneis :-)

I heard a great talk from Enrique Chaparro about security (he recommended python as a first language) and Lluis Sanchez about mono and monodevelop. Later I had a nice conversation with Lluis about the future of mono and gnome.

Later in the night I heard Richard Stallman’s conference about copyrights.

Today I finally gave my python talk with a nice audience>. Now I’m waiting for the meeting with the students that want to form free software groups or LUGs in their universities :-)

CONEIS stands for National Congress of System Engineer Students in Peru. I’ll give a Python talk this friday and I am also one of the hosts in the stand of the Peruvian Association of Free Software. Yesterday I met Alvaro Lopez, he gave a presentation about cherokee, a light and fast web server. I was very impressed and I’ll give it a look as soon as I can. In the night I met Rodolfo Pilas from Linux Uruguay, and later Diego Saravia from SOLAR. Later, Ernesto, Rudy and I had a nice dinner with Diego, and we talked about the possibility of working together as associations (APESOL and SOLAR).

Today in the morning I heard Rodolfo’s excelent talk about the Free Software’s Business Model. You can see some of his talks and conferences here.

We are the champions… my friends~~ :-)

Esta es la U:

\o/

Something scary hapenned to me this morning, my brain just stopped working correctly. After being awake for more than 36 hours and having more coffee than blood in my body, I could notice that something was wrong. It’s like when you are into some extreme and extenuous physical activity and there is a moment that your legs/arms just stop responding to you. That exact same thing can happen to your brain, and I’m telling you, it is not a nice experience. I couldn’t even configure a network interface in debian correctly. I was telling my brain “hey, we need to do that! we’ve done this thousands of times! proceed!” and the brain was like “huh? get some rest pal, I refuse to work any longer”.
Now back to work, it’s 4:20am and I still have work to do (oracle->postgresql fun). When will I learn? :-)

I always try to do the Right Thing ™, specially when the thing in question is job related. That is why I can’t understand how some people can take their work so lightly.
I’ve had a very, very bad time because a postgresql database collapsed, the reason being that the SCSI drive where the data resides had 0 bytes free left. I must admit that I didn’t know the internals of the custom-made-tomcat+postgresql-based-system in that server, so my approach was rather cautious. On closer examination of the box and after reading the “documentation” of the system in mention and the shell scripts written by the same author (those systems were deployed long before my arrival), I could notice that:

• The database had no VACUUMs in 2 years (broken scripts)
• The “cleaning up” of old unused data didn’t work (broken DELETEs)
• The backup procedure was idiotic: backup everything to a file, then copy that file to tape
• Since the database was HUGE (70GB), even a compressed backup would be a lot larger than the 2GB ext3 file size limit for files (see previous point). In fact, I could see the previous backups stored in /backup, and all were of the same size (2147483647 bytes). All the backups were broken and useless.
• Worst of everything, the design of the tables in the database was completly childish. There was one HUGE table with everything stored in there. There was no normalization (there should be death penalty for people that do that in a production server). Because of this, the application is very SLOW.

I also noticed that the custom-system is very poorly written. For example, after a Java procedure to insert data to the database, the user must READ THE LOG FILES and search for Java exceptions to MANUALLY insert the missing data in another table. This guy didn’t even know how to catch exceptions. I’ve seen his code in C, PHP, Java and shell, and it amazes me how somebody can be a bad coder in so many languages. I’m not talking about style here, but doing the Wrong Thing all the time.

At least the system in mention looks pretty, and that’s what managers will look at anyway. That guy, as a programmer, is an excelent graphic designer.

Oh, and the server is up and running again, I rewrote the backup procedures and I’ll replace that crappy system as soon as I can.

do it