Best Podcast so far : LugRadio

27 March 2006

I don’t have much time for this blog, but I can’t resist to say how I enjoy listenning to the LugRadio PodCast !
They are really funny and good sound quality !
Well, after 2 episods (happy pod) I didn’t learnt much technically I admit, but it makes me strive to impove my english, so I don’t miss a word !

LugRadio is a fortnightly British radio show that takes a relaxed, humorous look at Linux and open source.
Note: language may offend some.

iTunes shows a red “Explicit” tag on some podcast. I wonder why LugRadio hasn’t it, because they deserve it more than anything I heard.


MySQL Scalability

22 March 2006

I just read the Flickr architecture PDF by Cal Henderson .
It is not a perfect architecture, but still a profitable experience !

It is a bit old (2004), but I found quite funny what they did with MySQL. They started with one server using InnoDB (for transaction and performance due to row level lock instead of table lock). Then they replicated it horizontaly, meaning that only the master handle the update/insert/delete while the slaves handle the selects ! (Scaling vertically means that the application talks to all the servers : imagine that the app server box is on the left and the DB server boxes are on the right.) Then, they replicated some server but changed the slave’s table type to MyISAM (I didn’t know it was possible) so they can use FULLTEXT index, not supported by the InnoDB type. That’s tricky !

Appart from that, they confirmed some choices, that aren’t still obvious for everyone, like the use of UTF-8 and surprise me with the use of Smarty template, I thought that they would slow down things, but I guess the cache is good. I would like to read more details on how they avoid using sessions, but it is just a Diaporama presentation, so …


Java Posse Podcast

18 March 2006

I have been listenning to the Java Posse for a month now. Since I don’t have many poscast to listen to for now, I downloaded all the old episods. There are some pretty good interviews.
Over all, it is motivating and I guess that I will spend less time on TheServerSide, at least !
On the downside, they are too much about “Java is as good as Rails”, “Java is as fast as C”, as if people listenning to them, the “Java Posse”, didn’t already had an opinion … Then there is this IDE discussion that isn’t always covered by real information … and finally, they never resists to some speculation about what they read or heard in the blogosphere.
Well something more technical wouldn’t hurt. If they could try to get in touch with the people involved in the subject they talk about, it would be great. I think that Skype could allow some instant informations.
This is my opinion. Then, there is an article in ONJava by Chris Adamson about The Java Podcast. They state that there are only about 2 500 downloads which seems very few ! But I don’t know what is the average number for a Podcast. Anyway, I think that it will increase, like the mobile phones !


Poster for sale

12 March 2006

Some poster “à vendre”
Si quelqu’un est intéressé.
D’autres suivront.
Conrad ;
Spirou et Fantasio


Comparing Java webapp frameworks

11 March 2006

A work in progress by Simon Brown in java.net, heard from the Java Posse.
He is implementing the same webapp in a different framework each month. I hope that he will make some kind of conclusion to summarize his work.


WordPress 2.0.2 and UTW update

11 March 2006

I just updated WordPress 2.0.2 and Ultimate Tag Warrior 3.


awk, Linux utilities

9 March 2006

You must known the Linux utilities out there, they are really valuable !

I already know awk, but never actually use it until recently. I used sed, but not that much.

I didn’t see the real value of awk until I came to the following use case : a Word document to transform into a csv format ! I said that it was impossible : you have to to it by hand because it is unstructured !

Use the awk, Luke !

2 steps : save as plain text (note that it should be possible to do things with rtf too) Then write a simple awk script, like this one :


#!/usr/bin/awk -f
BEGIN {
    cat="categorie"
    FS="\\n"
    RS="Definition"
    ORS=""
}
{
    print  cat, ",", $6, $7, ","      # categorie, 6th and 7th lines concatenated, 

    # Skip lines to the real content
    x=9
    while ( $x != "Confirmation" ) {
        x++
    }
    x++

    # Print the content between Confirmation and Observations 
    # if it doesn't begin with a number
    while ( $x != "Observations:" ) {
        if ( $x !~ /[0-9]+./ )  {
            print $x, " "
        }
        x++
    }
    print "\\n"
}

Simple ? Yes. The important part is in the BEGIN block : awk cuts things in rows and columns. By default, it uses new lines to separate records and “,” to separate fields. BUT, you can change this : I tell him that the Field Separator is the new line “\n” ; the Record Separator is the word “Definition” and that the Output Record Separator is just a space. It is why I have to put a new line at the end.

My example may not be very explicit, but imagine that you can extract all the paragraph after the title “Introduction” or specifics parts of a document. It works great if your Word doc contains tables.

Here is a good references : IBM Common threads: Awk by example by Daniel Robbins. Check the part2, it is really usefull.