The theme is now built as a “child theme” of the toolbox theme, which spares me the burden of maintaining a lot of code since there is now hardly anything else to play with than the CSS file.

What’s more, this ‘toolbox’ theme leverages the new features of HTML5, so that the tlog is now cutting-edge (at least on the technology side)

This has also been the occasion to change quite a few visual details by playing a little more with rounded borders and small shadows, time will tell if it’s an actual improvement…

The theme can be downloaded here in its latest version.

Some resources on the same subject:

• The very famous extension Firebug, is still incredibly useful to play with themes without having to upload any file nor to reload any page in your browser.
• Good to know: the toolbox theme is not provided with the translation files (the .po and .mo that ought to be in the ‘languages’ folder) but it seems that the translation files for the TwentyTen default theme are compatible (and translated versions of TwnetyTen can be found via the codex).
• More information on ‘why using HTML5 for a blog’ in an article from Smashing Magazine.

By conviction and also because it’s just common sense IMHO,  I refrain from storing too much content (text, photo, code) directly into sites like Facebook (and as some would say “corporate silos”) that tend to consider their user’s data as their own a bit too easily.

More precisely, the best way to store ones data remains one’s own computer (with a tad of backup that is) and that’s the main way I store my photos and code.

For data a bit more “endemic” to the Internet (blog posts and selection of photos to be shown), I’m using “free” services built on free software and using open protocols that make it easier to connect several of them together. In this matter, WordPress, zenPhoto and identi.ca are doing a fairly good job for me.

Eventually for very specific cases I’m going through proprietary services like Flickr, Twitter and Facebook.

The aim

Well, the aim is quite simple at first sight: make it so that the new posts and photos published on my blog or on identi.ca end up displayed on my Facebook wall.

It should be “simple” since all the required technology exists, is free and utterly widespread with Internet and it HTTP protocol on one hand and on the other hand the RSS feed ubiquitous over the web and allowing tons of applications to “keep in touch” with the new content appearing on so many websites (new post, new photo etc).

The failure

Some time ago, Facebook offered to possibility to plug in several RSS feeds whose content would appear on one’s wall. It worked ok, and well enough to make a blog post about this pointless anyway. Since then, they limited the number of feeds that could be plugged to just one, which means that one cannot publish content from several sources and has to chose which one will get on facebook (either photos, blog posts, or dents/tweets). The most annoying thing being that even with only one feed the thing just didn’t seem to work properly.

So the message was clear: if you want to do such a thing, use a third party application. And I actually tried a renowned one supposed to do exactly what I needed  (twitterfeed) but that, despite having several interesting options to filter the feeds’ content etc, appeared to be nearly as bad at readeing feeds as facebook.

The solution

I eventually found another application that did exactly what I asked it to (at least so far): RSS Graffiti. This application transcribes the content of several feeds onto a facebook wall. Functionality wise that’s about all there is but its exactly what I need and above all: it works !

Going a little bit further…

For a little more refinements and to play a little with the feeds’ content (modifying the titles and filter some articles), it’s possible to use Yahoo Pipes: a very impressive tool which is also quite fun to use.

For those that intend to try, be aware that there is a drawback with using Yahoo pipes with an application like RSS Graffiti (the same would happen with twitterfeed and facebook): Yahoo sets a limit to the number of connection coming from a given IP, and as soon as too many people plug RSS Graffiti on their Yahoo pipes, Yahoo blocks (some of) the connections from RSS Graffiti. Result: RSS Gaffiti is unusable when plugged directly on Yahoo Pipe…

Happily a solution exists but then it becomes a little more hackish, even if there nothing there to impress anybody that has already played with a linux server for instance. What’s required is a server serving web pages and where you can set a cron job running wget to fetch the feed from yahoo pipes and save it in a directory accessible from the web. It’s the URL of the copied rss file that must be given to RSS Graffiti for the whole thing to work. Also note that one must be sure that the copied rss file is deleted before wget goes looking for another copy¹.

Update: Apparently the rss file generated by Yahoo Pipes does not declares its encoding in the xml header (the first line reads <?xml version=”1.0″?> ) but when several feeds, mostly in UTF-8, are joined into a single pipe and when somehow among all the content some non-utf-8 character appears this can cause some interpretation problems on some services (for instance accentuated letters are displayed as “weird” characters). A way to avoid these problems is to make sure that the UTF-8 encoding is explicitly declared, which can be done with a little play of echo and tails, so that the command would become:

rm machin_no_encoding.rss \
&& wget yahoo_pipe_url\&_render=rss -O machin_no_encoding.rss \
&& echo "<?xml version=\"1.0\" encoding=\"UTF-8\"?>" > machin.rss \
&& tail -n +2 machin_no_encoding.rss >> machin.rss

1. so the cron job is more like ‘rm foo.rss && wget yahoo_pipe_url -O foo.rss’

• serious blue tlog — this blog’s theme — now fully localised
• my Emacs configuration files (now with reST support)
• the todotxt Deskbar plugin and the todotxt Python module (with a few bugs corrected)

Du coup j’ai fait la mise à jour de WordPress dans sa nouvelle version (2.3) et en français. Le passage s’est fait sans douleur étonnement¹. Alors comme je voulais quand même souffrir, fo croire, j’ai décidé de traduire le thème de ce blog en français aussi. Le tout est dispo sur la page WordPress du site.

1. un bravo, donc en passant aux équipes qui développent WordPress et sa version française

Comme notre très cher président¹ propose de parsemer la France de caméras pour choper quiconque oserait se marrer dès qu’il apparait bourré à la télé, j’y vais moi aussi de mon appareil “tout sécuritaire”. Mais dans mon cas, ça se limite à mon blog.

En effet, je teste depuis quelque temps des plugins WordPress qui font des stats sur les visites du blog:

• Automattic Stats: dans la lignée de Google Analytics (que j’ai jamais essayé perso) il est assez sympa, les données sont présentée dans des graphes très potables mais il manque un peu de fonctions. En particulier, il ne montre pas la situation géographique des visiteurs et je suis pourtant fébrile de vérifier à quel point personne ne vient du Brésil, de Suède ou du Bhoutan.
• Quick Stats: bon c un principe similaire mais avec une présentation beaucoup plus spartiate. Mais bon il liste les IP et les noms des navigateurs Internet utilisés. Du vrai flicage donc, mais ça palie aux manques du module précédent :)

Un autre plugin sécuritaire me permet de filtrer le spam de façon assez efficace jusqu’à maintenant²

• reCAPTCHA: rien à voir pour le coup, juste la dernière incarnation des célèbre Captcha. Ces petites images qui présentent des lettres déformées et que l’on utilise pour savoir si la personne qui poste un commentaire est un humain ou un vilain robot spammeur. Si ce n’est pas très clair… z’avez qu’à essayer de poster un commentaire ici et vous comprendrez vite. Sinon ce que ce système de captcha a de particulier, c’est qu’il est utilisé pour aider à la reconnaissance de caractères dans des livres scannés, plus de détails sur le site[en]. Cette fois ce n’est plus du flicage mais de l’exploitation je sais, je sais… Mais d’abord je trouve l’idée marrante et, cerise sur le gâteau, c’est apte à énerver des collègues qui bossent dans le domaine :)

Mise-à-jour: on m’a fait remarquer que l’adresse e-mail était requise pour pouvoir poster un commentaire. Ca fait vraiment du flicage haute-dose faut avouer, mais pour tout dire je m’en étais pas aperçu (ben ouai je poste rarement des commentaires sur mon propre blog… J’ai donc désactivé le truc et j’en ai profité pour remettre un peu d’ordre dans les champs à remplir (ct le bazard). J’apprécie toujours que les gens qui postent des commentaires s’identifient un minimum (notamment le pseudo et le site s’il existe), mais c’est plus par curiosité et ce n’est plus obligatoire.

Mise-à-jour bis : J’ai viré le plugin qui servait à récolter les ip (entre autre), d’une part j’ai pas envie d’avoir de pb avec la Cnil et en plus j’en faisais vraiment que dalle…

1. vu l’état récent des comptes et les dépenses en vue, la vieille blague de Desproges est recyclable à peu de frais
2. je me demande juste s’il ne filtre pas aussi les “vrais” commentaires vu que j’en reçois quasi aucun ;)