Here is a quick list of improvements I made to the original b2 0.6.1 lately:

• New key features:
  
  • "View last comments" feature
  • Each post can be assigned multiple categories
  • Posts can be spread across multiple blogs
  • Statistics based on logging of Referers, User Agents, etc... on blog pages and RSS feeds
  • Comments, trackback and pingback can now be displayed on same page
• Backoffice improvements:
  
  • Improved standard markup buttons (support for strong, em, etc.)
  • All those little smileys can now be easily inserted by clicking them like a button
  • Possibility to send trackback and/or pingback as an afterthought (multiple trackbacks allowed)
  • Displaying progress while sending trackbacks, pingbacks and pinging various sites
  • b2 can now live in its own subfolder (does not have to be in blog root folder)
• Other improvements:
  
  • Performance: Reduced # of SQL queries per page
  • Performance: Optional generation of static home pages for each blog
  • Various bug patches
  • Standards: Various XHTML markup patches

I kind of started tweaking my b2 out of impatience waiting for a new version, and also because I thought it was an interesting way to refresh my PHP/mySQL knowledge... But this is obviously starting to get out of control! The time spent on this is begining to add up to something quite significant! :roll:

Thus, I am considering releasing this as open source. Though, some parts of the code are far from release condition... and if you have looked into b2 before, you may know the ground wasn't super healthy anyway :-/

So... if you are interested in this b2 evolution and you want to heavy-beta-test it, wether you've been using b2 or not in the past, please drop me a line ;)

Looks like mySQL is not able to to perform the DISTINCT selection in cases like this:

SELECT DISTINCT YEAR(post_date), MONTH(post_date) 
  FROM whatever1 INNER JOIN whatever2 ON field1 = field2
 ORDER BY post_date

This can be worked around, but it seems ridiculous!...

SELECT DISTINCT YEAR(post_date), MONTH(post_date) 
  FROM whatever1 INNER JOIN whatever2 ON field1 = field2
 ORDER BY YEAR(post_date), MONTH(post_date)

Actually I had not really looked into PHP since early version 3 betas... I knew it had gotten better, but sometimes I wonder if I didn't overestimate this... :-/

Looks like PHP 4.3 still requires that you name a collection of checkboxes or a select multiple with a name ending with "[]" such as <input type="checkbox" name="collect[]" value="3" />. That [] makes it impossible to DOM-handle the objects with javascript... but if you don't include the [], you just won't get all the checked values, neither as a table nor as a list... only the last one! :!:

What year is this? 1995? :'(

As far as I remember, ColdFusion and ASP have been handling this elegantly since version 1.0!

Ca faisait longtemps ça que ça ne m'était pas arrivé: ma carte bleue débitée deux fois du même achat. Il faut dire aussi que j'ai commis une légère erreur de client trop arrangeant! :no:


C'était un soir, dans un resto Montpelliérain, à la fermeture... je tape mon code machinalement, la serveuse reprend le boitier, il fait son petit manège habituel "brrr-brrr-tssst-bzzzzt-bip" mais il ne sort pas de ticket! "oulàlàlàlà j'ai un problème de machine" dit-elle avec un sourire gêné de circonstance... "il n'y a plus de papier" et zou elle disparait... au bout de quelques minutes je prends quand même doucement conscience que j'ai laissé filer la serveuse avec ma carte... que je ne sais pas ce qu'elle est en train de faire avec et que -- si ça se trouve -- elle m'a regardé taper mon code! (mais non je ne suis pas parano :roll:)


Bon, là je me dis qu'on va pas me la faire à moi et que je vais pas me faire avoir comme un bleu. Nous nous levons donc et arpentons le restaurant (qui se vide de ses derniers clients) à la recherche de la serveuse... enfin surtout de ma carte! :-/


En fait, on la retrouve avec le directeur, lequel avait étalé sur le comptoir... (mon Dieu, en écrivant ça je me rends compte qu'on a l'impression qu'il va se passer quelque chose de terrible!!!:D)... sur le comptoir donc, tous les tickets de CB de la journée ( ;) ), calculette à la main, en train d'essayer de recalculer la différence entre la caisse et les machines à CB... L'objectif est de me montrer qu'il y a un différentiel de 22 €, donc que ma carte n'est pas passée... Manque de bol, après 30 minutes de calculs (non non on était pas pressés), il n'arrive pas à descendre en dessous de 150 € de différentiel... Malgré ce chiffre hasardeux, je me laisse tout de même convaincre sur le fait que "c'est pas passé" et je retape mon code une seconde fois. Erreur! :!:


=> Règle n°48 de l'utilisation de la carte bleue: ne jamais retaper son code si on a pas un ticket sur lequel il est inscrit en gros "ANNULATION"!


Parce qu'après... il va en falloir de la persévérance de principe, pour récupérer ces 22 malheureux euros!

Post complet »

As Russel points out, the quality of searches is dwindling at Google as the result of BlogNoise.

It seems to me that Google could easily cut out a lot of blogcrap of their search results if they performed their searches on a post by post basis (all words have to be found in the same post) instead of a page by page basis (a weblog page contains an average of 15 very loosely related posts).

How would their indexer find out about the boundaries of each post on a page? Well... just let it take advantage of the RSS linked to any decent weblog!

Oh wait... Blogger blogs don't have RSS! They have a huge market share (i-e a huge blogcrap share) and if they still haven't implemented such a straightforward feature yet, they're not very likely to do so soon... That's a problem!

Let's acquire them and show them the way! :>>

Semantic web rules!