Month: November 2006

Ruby France logo proposals

I couldn’t find sleep tonight (*) so I did two small variations on the Ruby official logo for Ruby France, since they are looking for a new logo (**). Double-clic to enlarge, single-clic to reduce back to the small images (***) :


I also like Greg’s proposal.

(*) it happens very often these days
(**) No, I don’t know Ruby
(***) Doesn’t work if javascript isn’t enabled in your browser (usually it is) + Internet Explorer doesn’t correctly render the transparency ; please use a real browser instead

Rose Annual Dinner

On Saturday, we went for the Rose Dinner 2006. Rose is an independent not-for-profit organization run by volunteers. Its mission is to support and foster primary education for children from economically disadvantaged rural communities in developing countries (mainly India), primarily by improving basic facilities in primary schools. Yesterday, its impressive team of cooks offered us great food (thanks Swapna), there was a good atmosphere. It’s good to see that people that actually do something in an effective manner. 🙂

We also met Peter (and his Dutch friend Peter) and, for once, we had some time to listen to his funny stories about some of his travels in India. And no, these stories don’t remove your will to travel to India (in fact, we are more eager to go back there). It was very nice to meet them.

And, finally, ISAL Christmas celebration will be held on December 16th (at 19.00), in Leuven. Save the date!

The Digital Ice Age

In this article, Brad Reagan gives many examples where the use of electronic data begins to cause problem, in a preservation perspective. The causes can be a new software that is not fully compatible with previous data models, new physical formats (unable to play old formats), too much raw information, etc. For the moment, free projects like the Internet Archive or the Free Archive (a.o.) are trying to cope with this problem.

Although the dangers of a “digital blackout” really exist, I think the author forgets one important aspect of information from the past: we already lost a lot of it. What is left is what time left us, often with some damages. It survived time, taking many different forms and paths, different storage procedures, different media, different locations, etc.

Instead of trying to store everything, maybe we should look at storing the most relevant information only. But now the question would be: how to know what is relevant and what is not? I vaguely feel that one can also add a notion of time: the e-mails I receive about a future party or a new product are not worth for more than 1 or 2 years although electronic exchanges that led to the discovery of siRNA, e.g., are much more important.

Plugins for Digital Object Identifier lookup

I’ve just written some “search plugins” for Firefox (1.x and 2.x) that allow you to quickly look for a specific Digital Object Identifier (DOI). These DOI are more and more used in biomedical sciences. One of their interesting features is that they allow direct linking to the scientific article.

The plugins are availble here. If you already have Firefox 2, the installation procedure is very easy: all you have to do is go to the plugins page, click on the small arrow near your Firefox search box and choose the “Add DOI lookup” option; it will then automatically be installed for you.

How to remove files ending with '~'

The vim text editor always produce a file ending with a tilde (~) as a kind of backup of the currently modified file (this is a default behaviour). On my MS-Windows machine (Pentium M, 1.73GHz), I was tired of manually deleting these files so I first used the “Search” option in the File Explorer. After some time, I got tired to wait for the results.

So I wrote a Python and a batch scripts to find all these files. They are going much faster than the Search GUI. The first time I launch them, they are still going slow (but faster than a GUI). As you can see in the graph below, the second time I launch these scripts, they went at least 10 times faster. I’m not a specialist but I guess it has something to do with caching at the OS level. For the first run, the batch script is 20% slower than the Python script. After that, the Python script is 50% slower than the batch script (but between 3.7s and 5.6s, the difference is not big).

Comparison of .bat and .py files

Here are the scripts : find files ending with ~ in Batch (the problem is that you have to do the duration calculation by yourself), find files ending with ~ in Python and remove files ending with ~ in Python (all scripts are 1kb).

Each of these scripts were run as the first application after my computer was turned on. I didn’t repeated the measures (doing real stats wasn’t the goal anyway). Deleting all the files (after having found them) took 5.4s. It just goes to show what we can do, just before the beginning of a lab seminar.

OSS/FS players about GPL Java

Sun opened Java in the most elegant way of doing it (imho): the licence is the GPL. This move was analysed and commented by many people. Even some important Open Source/Free Software players gave their comments on a Sun website. Unfortunately, their comments are only available in a proprietary video format.

You can now have access to audio recordings of these interviews (Brian Behlendorf, Paul Cormier, Eben Moglen, Tim O’Reilly, Mark Shuttleworth, Richard Stallman and Dr. Marcelo K. Zuffo), to a text transcript and even to SHA1 sums of the audio files!

Double quotes!

GGGRRrrrrrrr … I was quietly using R to analyse my data when, suddently, I wasn’t able to open the file containing these data anymore. It’s just a plain text file! How can it be corrupted? Here is the error message:

t < - read.table('ratsdata.csv', header=TRUE, sep=",")
Warning message:
incomplete final line found by readTableHeader on 'ratsdata.csv'

For hours, I tried everything: I counted the number of separators on each line, I counted the number of decimal points on each line, I removed double quotes around factors, I examined in details the final line, etc. (well, Python scripts did the job for me because my file already has > 700 lines). Finally, the solution was so dumb: I mistakenly deleted one double quote before a header. My first line looked like (1) and should look like (2):

1: "id", "group", "trial", "durTot", durExt" [...]
2: "id", "group", "trial", "durTot", "durExt" [...]

Ok, next time, I’ll add this check on my list …

White & Nerdy

It’s Sunday, let’s rest a little bit … I really liked this Al Yancovic‘s video “White & Nerdy“. To fully understand it, you need some basic technical background and a friend that looks like the white & nerdy guy in the video. Because, of course, you are not like him 😉

Dont’ download this song” is also great (some background about DRM is welcome).

It’s all about Pentiums” is not my music style but some lyrics are good.

If you don’t like Flash movies, you can use the Firefox OOk video plug-in to obtain the FLV file and then read it with VLC or convert it with ffmpeg (e.g.).