filmo (animeo): somera batalo

Pasinta semajno mi vidis tre agrabla kinejo, estas nomiĝa サマーウォース (summer wars).

Se tio kinejo estis menta por esti en la 1900 aŭ 1700, la personoj de tio filmo povas partopreni en batali kun armoj, sed, la ideo de tio kinejo estas la "nova batalo", kie estas en "virtuala mondo" (kiel facolibro aŭ tŭitero)

Mi pasis tio filmo (kun angla subtitulito, mi ne povas kompreni la japana lingvo pli bone) al amikoj, kaj li pensis ke estas tre bona kinejo ankaŭ.

Kiam ni vidis tio kinejo, unue mi pensis ke tio estas komedio/amo kinejo, sed tio havas amo kaj humoro, sed estas epika kinejo.

Alia kinejo ke mi vidis lasta tempe kio estas pli agrabla:

• Time of Eve
• Wellcome to the space show

Kaj alia animeo ke vi povas vidi (kun esperanta subtitulo) estas Mia najbaro Totoro

HP Blade and gentoo memory leaks

Hello minna san!!

I had some terrible troubles last month within two host that I have with Gentoo over HP Blade 460c G6. The problem was simple, the most IO that I generated in the host, the more memory in SUnreclaim (in /proc/meminfo).

The problem (after a log of time and testing), was solved by configuring the kernel with:

# CONFIG_IOMMU_API is not set

And, after a whole night of testing (that in prior test sessions in 2 hours the memory claimed up to 3 or 5 GB), was just in 80 MB the whole night.

The problem is quite known with the iommu module, but at first time I thought that was the cciss module (because everything was related with the disk IO). But after a further analysis, the iommu came across my path and then I saw a lot of bugs related to this module (and most of them were related to the memory leaks).

hosiaux gentoo kaj ruby (rubio)

Hosiaux mi estas tre tre laca, mi dormis bone, sed mi estas laca. Mi debas instali HP Blade kun gento (tio ne estas problemo, sed la HP Blade, kiam mi re-startigis la sistemo, dauxrigas multe tempo.

En tio tempo, mi volis scribi en mia twitero, sed en mia laboro ne estas pli bone uzi twitero, kaj por tio, mi kreis simpla rubio programaro:

#!/usr/bin/ruby
require "rubygems"
require "twitter"

Twitter.configure do |config|
  config.consumer_key = "something here"
  config.consumer_secret = "somethig else here"
  config.oauth_token = "yes, other thing"
  config.oauth_token_secret = "the last secret"
end

client = Twitter::Client.new

client.update("my first ruby.twitter test")

Mi preferas Perl, sed Ruby estas tre amuza por mi.

Rosario

Rosario

syslog-ng durmiendo

Syslog-ng

Syslog

Rip! esta foto es del 2007 01 13

Donde trabajo ahora

Desktop pet

Alice

my new project

For a long time I was a bit depressed, mainly because I quit my past job and the promise of a new job was dropped. And then I couldn't get a job (the interviews that I had were't good enough to me). But since July, I'm working again. The job is quite good, I have some regrets and complains, but I enjoy it (by now).

For a couple of months I'm working (not full time) in my japanese-english dictionary (web interface). The goal of this is to let some of my classmates (I'm in 2º year of japanese) to have a nice japanese-english dictionary (the actual dictionaries available are quite disgusting -lack of vocabulary-).

And other part of my project is to learn a new framework. I tested ruby on rails and others, but by now, I'm quite comfortable with Jifty (perl based MVC). The reason is that I love Perl, and, with few code, let me do some things that in ruby on rails and others, I'll have to change the default behavior, and that is quite annoying to me. But the main reason to avoid ruby on rails, is because my last project went really disgusting with that framework (a web cook book, I had some disgusting problems with the presentation of the recipes), so I'm trying to avoid it.

Now I have a local SVN, when I could everything in order, I'll put the code in google code or some other site.

The goals of my project are:
* Simple web interface (few buttons, simple list -word;definition;dict name;word type-)
* By having a word type column, I could make a filter by adj-na, adj-i, noun, whatever (really useful when you are studying Japanese).
* A radical search form (this is quite simple, I get a list of Kanji-radical, so I can make a list of radicals and search by that list in the kanjidic file/db).
* A contributed part (a kind of user comments).

Other improvements that I was minding couldn't be done (like year/level of each word), because there are many books and level definitions. And I don't have any help of a Japanese linguistic (that could give me that list -or at least, the most common word level classification-). But I think that that is quite difficult to be done (or agree with that level classification).

By now, that I did was:
* dictionary normalization and import in an SQLite3 DB.
* web interface: word list and simple ABM to add words (but I'll put this offline in stable release, or make my web with some level of authorization).

I'm having some problems with the framework, because I'm not used to code in it, but is a matter of time while I read the documentation and I get used to it.

If someone want the code, just mail me (but is quite useless right now).