Here is my problem. I set up monitor.vrypan.net using lilina. It is actually an aggregation of many Greek blogs. What I want to do is to provide a “common place” for greek blogs to interact with each other. So this is not “just a blogroll”, it is (or should be) something more.

This was a good oportunity to experiment with an idea I had for some time. I used the Technorati API to check how many blogs link to each post, in order to “illustrate” the “virtual conversations” between blogs.

I then came up with an unexpected problem. Some blogs use Feedburner (an excelent service btw) to export their feed. This is a good idea, since Feedburner provides nice features and reduces the traffic caused by feed readers and aggregators. One of the coolest features of feedburner is tracking of clicks: Feedburner will not expose the original post URL but a custom one that redirects to the original. When a user clicks on it Feedburner will track the click and redirect to the original URL. So, using the technorati API on “fake” Feedburner URLs is useless as they are not “real” blog URLs.

I think that there should be some other way to do this. A possible solution:
1) Feedburner includes the original URL’s MD5. (i.e. instead of http://feeds.feedburner.com/Vrypannetweblog?m=34 it will show http://feeds.feedburner.com/Vrypannetweblog?m=34&md5=xxxxxxxxxxxxxxxxxx)
2) Technorati will allow the use of MD5 instead of URL.

Is this possible? Do you think of some other to get arround this problem? (Sidenote: MD5(URL) can also be used by del.icio.us)

A shinny new version of lilina is out!

0.7 has many improvments, here are the latest entries at the Changelog:
* Update cache only if ?force_update=1 is passed as a parameter.
* Cleaned up the grouping
* Expand/Collapse changes the color of the item.
* Initial Expand or Collapse is done by JS
* UI is much simpler (to allow easier customization)
* cleaned up a lot of things
* “Add” button in edit.php
* MAGPIE_USER_AGENT is set to “lilina $LILINAVERSION (+http://lilina.sf.net/)”
* inc/chkenv.php will automaticaly do some system checking (directory and file permissions)
* Should generate Valid XHTML 1.0 Transitional (depending on the contents of the feeds)
* CSS cleanup and improvements
* ALternative CSS (style_fancy.css) illustrates how look can be changed.

code.google.com: Welcome to Google Code, Google’s place for Open Source software. Read the FAQ for more information.

Nice!

I finally decided to move my greek blog from drupal to wordpress. Since there was no migration script, I wrote a couple of sql statements that moved all posts, comments and categories from my drupal tables to the (new) wordpress 1.5 tables.

Here is the proccess in short:

WARNING!!! This may delete your DATA!!! Make sure you backup EVERYTHING before starting the procedure!!!

1. setup a fresh wordpress installation.
2. make sure term_data term_hierarchy node term_node comments (drupal tables) are in the same DB you use from WP.
3. Run the following SQL statements:
delete from weblog_wp_categories ;
delete from weblog_wp_posts;
delete from weblog_wp_post2cat ;
delete from weblog_wp_comments ;

INSERT INTO weblog_wp_comments (
comment_post_ID, comment_date, comment_content, comment_parent
)
SELECT nid, FROM_UNIXTIME(timestamp), concat('',subject, '<br />', comment), thread FROM comments ;

You should now have all your posts and comments and categories in WP. Go to the admin interface and make sure everything is in place…

Notes: This is not the perfect way to migrate. Comments are not nested in the right way. A lot of things may not work. On the other hand if, like me, made a really simple use of Drupal, this should move most of your data to WP…

Ok, this is one of the things you come accross and say “Wow”! This guy, Amit Singh wrote some piece of software that makes use of PowerBook’s motion sensor. You can actually tilt the laptop and navigate a web page, or do other cool things!

Read all about The PowerBook Sudden Motion Sensor!