Interview at Google

This Wednesday I had an onsite interview at Google’s lovely engineering site in Zürich. It was a long session with different people interviewing me on various subjects. The interviews were engaging and interesting, and I admit I had fun. A lunch pause was accorded where I had a nice chat with one friendly mountain-loving googler, and I had the opportunity to taste good food at the restaurant. It seems to be the best workplace there is. Everyone seemed to be having a good time. Reading about it is one thing, but seeing it is something else. Lucky googlers.

Windows 8 will be unable to win against Android

Windows now has to compete without the network effect that kept it being the most popular operating system. The ARM tablet world already has its operating system, it’s called Android. The Windows tablet world is basically nowhere, and Microsoft is coming up with a new operating system where the classic Windows apps will not run (in ARM). This puts Microsoft in a situation it hasn’t been since decades ago: they have to compete with the incumbents without the network effect in their side. They will not be able to convince users to switch to Windows 8 tablets instead of Android or iOS ones through the argument of legacy application compatibility. It needs to compete bringing a new OS that has virtually no applications in an environment where Android and iOS already have an established application ecosystem.

In the Intel world Windows can still run most classic applications, but even then they look ugly and run clumsily compared to the new Metro. The experience will be one of very little consistency. I suppose most people will want to ditch the old apps as soon as possible. Even then, Microsoft insisted in using the old-style interface for some of the pre-installed applications, like Windows Explorer, with the aggravating (for some) addition of the Ribbon everywhere. I concede nonetheless that Microsoft is in a better position in Intel platforms, but the Intel architecture itself seems to be in a bad position. Higher power consumption (low battery life), slow start times, more heat, and no always-on philosophy in machines that become smaller and lighter, means that Intel processors lose appeal.

I predict that Microsoft Windows 8 will not succeed reaching even half of the tablet market share, and that desktop and laptop computers will be slowly decline in sales than ARM tablets over the course of the next few years.

Moving to ARM

I’ve decided to get myself an Asus Transformer Prime in order to code personal projects while on the go. I wasn’t expecting too much of it, but only the capacity to run a terminal emulator and SSH to connect to my home machine and code. But the quality of the machine, battery duration, connectivity,, Linux, and the fact that it is always on, makes this the best computer I’ve had so far. I’m now moving my computing to ARM as my main platform. I got root access, installed the Ubuntu text-mode userspace and tools, and made it work very well for Node.js, C/C++, Perl, Python and Scala development. If my exigences as a developer are catered by this little Android beast, I expect there will be an exodus from Winntel around the corner.

Bullshit Buster

If nobody is up to the challenge of removing laws against criticizing, mocking or insulting, then I’d propose we replace these laws for a single law against the dissemination of false information. So, nothing that isn’t proven true can be taught as such, and can only be accompanied by an appropriate disclaimer of conjecture or fiction. I guess all religions will be crushed pretty quickly then, for the betterment of everyone.

Stop SOPA

In about eight hours, the Wikipedia website, as long Reddit and many others, will join in an unprecedented effort to stop a dying industry from hijacking the freedoms of American citizens. The intellectual property wars is escalating, and it can only become harsher if the governments of the world don’t understand that the most common forms of intellectual propery (patents and copyrights) are obsolete and should be entirely removed from law, as evidenced by Rick Falvinge in his blog. The intellectual property wars will not end until that happens. What’s left of our right to share and create are in great peril. Let’s not give up these freedoms to corporations and governments as we have been doing in the past. There’s no place for patents and copyrights on the Internet.

If you’re thinking that you’re safe from laws like SOPA and PIPA because you’re outside of the USA. Think again, many of the laws that your country abides by are pushed from Washington.

Ubuntu Private Folder

I’ve been using full home directory encryption in my Ubuntu system for awhile. Although it works very well, there’s a slight performance impact. Not a big deal, but still, I don’t need to encrypt everything in my home directory, just some sensitive stuff. I decided to decrypt it, but still use an encrypted «Private» folder for sensitive files.

To create a Private folder, I used the ecryptfs-setup-private command. This was fairly straightforward. The interesting part came later: using symbolic links to allow applications to find sensitive data in the Private folder.

First, let’s move the Firefox passwords:

mkdir ~/Private/Firefox
mv ~/.mozilla/firefox/*.default/{key3.db,signons.sqlite} ~/Private/Firefox
ln -s ~/Private/Firefox/* ~/.mozilla/firefox/*.default/

Next step, my Bitcoin wallet, stored in ~/.bitcoin/wallet.dat:

mkdir ~/Private/Bitcoin
mv ~/.bitcoin/wallet.dat ~/Private/Bitcoin
ln -s ~/Private/Bitcoin/wallet.dat ~/.bitcoin/

Finally, the GnuPG keys:

mv ~/.gnupg ~/Private

Other sensitive files that aren’t related to any particular application I can also place in the Private folder to protect them from eavesdroppers. Now I don’t have to encrypt my whole home folder to protect a few sensitive files. In any case, full home directory encryption works very well in Ubuntu, and is the fastest encryption overlay I’ve used.

Desconfía de los psiquiatras

Un psiquiatra decidió poner a prueba los diagnósticos de trastornos mentales enviando gente sana a una serie de hospitales psiquiátricos en EEUU en 1972. Los pacientes debían actuar normalmente, como gente sana que eran. Sin embargo, debían decir que sufrían de ciertas alucinaciones auditivas, lo cual por supuesto era falso.

Los hospitales no lograron identificar a los pacientes falsos, asignando diagnósticos de esquizofrenia o desorden maniaco-depresivo. Las estadías fueron de entre 7 y 52 días.

En una segunda segunda fase del experimento, se les indicó a los miembros de un hospital que les serían enviados pacientes falsos, y que tenrían que detectarlos. El hospital alegó que 41 pacientes eran impostores, y otros 42 eran sospechosos. Lo cierto es que todos eran pacientes legítimos. Los investigadores no enviaron ningún paciente falso al hospital.

Los resultados del experimento se publicaron en la revista Science, dejando a la práctica del diagnóstico psiquiátrico en descrédito.