Older blog entries for Akira (starting at number 72)

16 Sep 2007 (updated 16 Sep 2007 at 09:20 UTC) »

Finally, I am there! the coding session started.

I received the lost book yesterday morning and I have already finishes this second reading (first was during my thesis given up in 1994 to focus on my very first startup). By bits, I remember the explanations of the mechanisms. A true pleasure.

So, I started to code these mechanisms in Python by using the same approach as the author : implement a very basic filtering engine then incrementally add more complex features (kind of roadmap) :

  • filter0: the very basic filter,

  • filter1: with a catch-all filter,

  • filter2: to be able to filter a filter (the lisp quote),

  • filter3: a data driven implementation,

  • filter4: filter but also extract the data,

  • filter5: filter segments of data,

  • filter6: use explicit continuation to manage nondeterminism,

  • and so on ...

The final version of this work will be branded "pyfilter". Parking to sourceforge dot net with a GPL v3 license is envisaged.

For now, this experimental development can be followed here. Do not hesitate to make comments or proposals, the long term objective being to obtain a good usable and used library to filter information.

Coding Vacation

A few years ago for my thesis, I was interested in the implementation of engines to filter information - using Lisp language.

One of the sources of inspiration was work of Christian Queinnec and in particular its book “filtering: an application of (and for) lisp” unfortunately available only in French.

I have just ordered again this book (lost in my many relocations) and I should receive it soon.

Indeed, I intend to take one week of vacation to code! and I envisaged to produce a generic engine to filter information - engine written in Python.

Soon or later, this engine will be used in the iTrade project to extract financial information from websites ; but I intend also to comment on and to develop this library to improve the effectiveness of Python when it is a question of extracting and managing information.

I wish to show that the power of expression of a filter makes it possible to appreciably reduce the size of the code necessary to extract information and that it is possible in Python.

Not to waste time for the logistic of this new project, I will park this work under a sub folder of the iTrade project. I would give more information later on when the coding session starts.

NB: for the name of the project, I imagine something around “Paragon”.

Some politics

    As in a good fable, the electoral promises respected or not, always finish by a moral.

    We arrive soon at this moral : it is not enough to launch simplistic slogans and ideas (i.e to work more to gain more), or to exploit the credulity of the voters using noisy media campaigns, to regulate the problems.

    The economic and social situation in France is in sorry state. It is necessary to be serious to face this reality.

    Our hyper president Sarkozy will not be able any more to make us believe that it acts because it is shown on TV.

Dear cat,

Many new users have empty fields, without personal information and without any attempt to login :


Member since: 2007-08-27 16:14:56

Last Login: N/A

No personal information is available.

Is it possible to filter such user creation ?
ncm , I've just added the advogato rss flux news to the projet.

Donate your opinions and save the planet. here.

“Choose a work that you love and you won’t have to work another day.” - Confusius

Burn Out

Just read an interesting article about how not to burn out yourself nor the people working for you.

Given that the Happiness equals the Reality divided by the Expectations, if your reality is too poor or if your expectations are non reachable, your Happiness goes near 0.

Finally, you have only two solutions : decrease your Expectations or increase your Reality.

Personally, I prefer the latter.

63 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!