zengun

Following one of the current trends (I can’t afford following every trend under the sun, you know), I’ve been thinking of adding a “List o’ Links” (like Anil Dash‘s, to link a popular example of such a list).
After all, sometimes I find stuff that I want to blog about, but don’t find the words to make it a blog entry or write an half-assed entry about why you should really go there. I’ve been known to even resort to the ‘nude 24/7′ sellout trick to make readers click on a link.
So yes, I could do that. I first need to redesign, I’m growing tired of these collapsable sidebars that maybe no-one uncollapses and that hide content that could otherwise be straightforward visible.

And then, I thought it would be cool if I could just modify my command line blog client to make blogging a link as fast as bloglink http://someURL.com, the blog client automatically fetching the URL to get its <title> and building <a> link from it.
It was all fine and dandy, until as a test I wanted to blog about Dorothea Salo’s entry on Blogging for dollars, and ended up with a link that read “Caveat Lector: Octubri 13, 2002 – Octubri 19, 2002 Archives”: now, what a descriptive title for a link, I can already hear you thinking out loud “THIS IS SO DESCRIPTIVE I’M GOING TO CLICK ON IT TO FIND OUT MORE”. Yeah, but… not.
This is not a problem with Dorothea’s permalinks, this is a problem with how the app was thought out. So instead I’m going to modify the blog client so that it takes a description argument, for I’ve learned I should never assume every URL to have an unique <title>.

Or should I? Wouldn’t it be a good idea to append the <a>nchor name/id attribute to the title in the browser’s titlebar and maybe in its document.title counterpart in the DOM? The latter would be too dramatic of a change, so I’d just forget about it, but how about the titlebar?

Anyway, all this rambling talk, and I still didn’t blog these links:

• Dorothea Salo on Blogging for dollars and how it can’t be harmless to a blogger’s integrity.
• Simon Willison on why prefetching link=”next” URLs is everything but a good idea.
• Big business has no compunctions about changing the ground under us, so why not do it back?

I’ve just added a cool feature to the way b2 outputs data: an optional $b2_filter array that I fill with template tags names and associated functions to use on their content.
To put it in simple terms, let’s say I bang my head on the wall and come up with the crazy idea that I should display an md5 hash of my titles instead of the regular titles. I just have to add this to my config file: $b2_filter['the_title'] = 'md5';.
It doesn’t stop at just one function, I can apply several functions to a template tag, like this: $b2_filter['the_content'] = array('md5', 'strtoupper');.
And I could just say all template tags that support this filter should display as uppercased md5 strings like this: $b2_filter['all'] = array('md5', 'strtoupper');.

So this allows me to apply Matt’s PHPCurlMe function to my content, which is a lot less messy than hardcoding these smart/curly quotes before posting an entry.
Then again, now I have to modify the function so that it doesn’t convert smart quotes in text that’s between <code>, <kbd> and other similar tags…

Just days after the new release of MT, Hixie’s Pingback vs TrackBack whitepaper is back for the TrackBack 1.1 spec (which has pretty much turned into a Pingback-wannabe spec, except for the crazy idea of auto-discovery by parsing RDF in comments).

The more I read the way this courageous Python wrestler hacked TrackBack into Radio, the more I smile at how it would be so easier to achieve the same thing with Pingback.
Also, ALA shows you how to do a PHP stylesheet-switcher, probably what, one year after I used that almost-same code on one of this blog’s earlier designs. It’s nice to know that had I known this code could be featured on ALA, I would have written an article about it and gotten some exposure. Actually, I now have an idea of an article to submit them, wish me luck to get it published.

Wired.com redesigned, and the least one can say is that some company in Redmond should have hired these designers for their redesign earlier this month.
Wired not only validates as XHTML 1.0, but it’s also browsable in text browsers, with navigation help and all goodies. (via Simon Willison)

This morning, I went to the unemployment office again, only to find that I missed two more papers. It’s embarrassing.
It’s embarrassing to find how untidy I am when it comes to my identity. I drop bits of myself everywhere I go (Vallecalle, then Montpellier, then Cervione, then Corte), and then forget to pick them up before I go: even two years after I left Montpellier, my social security number is still handled there, because I forgot to tell them I moved back to Corsica. I left Cervione and my previous job in a hurry, and forgot to ask them a paper stating that I really worked there and for how many hours.
I lost my cellphone in July, and some lost friends with it, since I lost their numbers and they didn’t try to contact me.

I read Gunnm’s 8th volume once (Battle Angel Alita, for americans), and there the crazy scientist character who died and rose back from a backup of himself. Said it was always a good idea to have a backup of oneself in a secured place.
I’m starting to wish I had made backups of myself, and scattered them all around. Just to make up for that constant eroding of myself, coming with my every move.

I “hijacked” John Gruber’s Educate Quotes module, and made it into a stand-alone Perl script and it works — “Smart quotes”, at last !
So now, I can just write entries in a text editor, save as ‘newpost’, use the Perl script, and use blog2 to post the file as a new entry.
I might even just add a system() call in blog2 to get the results of the smart quotes converter, so that I’d just have to post normally with blog2 and not care about converting stuff. Oh, and I shall just convert the Perl code to PHP, and integrate that into b2 as a display option. So I wouldn’t have to care about anything once I set the option on. Later, later…

‘Ah’, “with” ‘these’ “quotes” ‘I’ “feel” ‘the’ “power” ‘of’ “ten” ‘Dean’ “Allens” ‘flowing’ ‘through’ “my” ‘fingers !’

While driving, one should always stay away from the mud.

It‘s a meme revival !
B7 d++ t++ k+ s+ u f+ i- o x e+ l c

Check this out before the comments get edited or Brad Chaotes’s MT Sanitize plugin gets ‘fixed’ to avoid CSS defacements: OMFG, I just 0wned Phil Ringnalda !!!!!111 (screenshot)
CSS can be so fun :P

Update: actually, Brad already fixed that in his plugin, it’s just that Phil was using an older version of it.

Oh, but I forgot to blog this little piece of PHP that I finished some days ago.
I guess this won’t be of interest to much readers, but here’s what it’s about: you give it a dictionary file that has undergone affix compression and its related affix file, and it will inflate the list into a simple words list just by reverse-applying the affix rules.
For my experiments, I choose the English and French dictionaries from OpenOffice.org. If I remember correctly (CBA to test again, sorry), the English dictionary inflated to about 150000 words while the French one inflated to about 650000 words. The difference is mostly because of the prefixes abundance in French: since prefixes get to be applied suffixes too, they tend to make the words list huge. I can’t fathom the size of the inflated German dictionary…

Some good news:
- I finished my PHP command line Blogger API client, it’s there (readme is there) and it’s now called ‘blog2′. It’s released under the Artistic License (I wanted to use the GPL, but then I couldn’t merge IXR
right into the code because of GPL vs Artistic License conflict).
- Google finally re-indexed me ! I noticed that when I saw someone searched for “gaim trillian buddy xml”, and sure enough I chatted with the person and he could import his trillian buddy list in gaim thanks to yours dearly.
- I have finally found a name and domain name for a blog in french. I was surprised the domain name wasn’t already taken, actually. More about that later.

One thing I find odd, with the few command line blog clients that I had the chance to try, is that none of them included a way to just append a string to a post. For example, to just say “Update: it works now !” you’d have to edit the whole post.
So I added this possibility to my client. Now I can type b2 g last10 to get a list of recent posts’ IDs and excerpts, and then say I want to add an update to post #79 I can just type b2 e- 79 " Oh, I forgot to say today the sgurz was great !". I can also type b2 -e 79 "Today was fun: " to prefix the entry with that statement. And then to get fruity I could type b2 -e- 79 "***" to enclose the entry between stars. It also works with the ef (edit from file) option.
Of course, this still needs to do a blogger.getPost query, but it’s not a biggie compared to the benefits, is it ?

Obviously it’s an ironic metaphor for the transubstantiated existence of the post-blog phenomenon and – indeed – the meta-blog in it’s self, a cunning statement on the semantic web and it’s innate pointlessness in a society of mere opinion and painfully little content with the post-metablog schematics linking into a new literary style of personal ‘push button’ publishing, sending the traditional model of editorial spiraling into the plasma explosion of a new dawn.

Or it could be that I have just got the Blogger API support working in epistula.

Looks like I’m not the only one playing with IXR and the Blogger API ;)

I’m having some coding fun with IXR. Simon Willison really made coding with XML-RPC a piece of cake with IXR.

So my quick command line blog client now uses IXR, and is almost completely functional. It posts here.
I can already issue commands like: b2 p 'this is my post to post short entries.
Or I can blog a file, like this: b2 pf mypost.txt to post the content of mypost.txt.
Or I can get the last 10 posts like this: b2 g last10, that will list the last 10 posts with their ID and an excerpt of their content.
Then if I see post #14 in this list and I want to edit it I can just type b2 g 14 >post14.txt and then edit post14.txt in my text editor. And then I’ll post it back with… wait, I have to code this function. Back to the drawing board to finish the client !

After checking Leonard’s compilation of HipTop’s barbarian crimes against the web, I wonder how tidakada.com looks like on this HipTop device. And I’m shivering.
Why do I have the feeling that the CSS positionning used here would be very, very ackwardly interpreted ?

In other news, I took a big dose of geek courage, and actually set up a new mini-blog right from my domain’s SSH shell. I even edited a config file with vi !
But my other geek mission, writing a quick command-line blogger-api client in PHP failed. Anyone who managed to get UsefulInc’s PHP XMLRPC classes to work in command-line PHP is invited to show me code, for now all I got is ‘Cannot instantiate non-existant class: xmlrpcmsg’.

And oh, a quick trip to Simon’s tells us Sam Ruby’s got pingback. And you, got pingback yet ?

I re-installed Linux again two days ago, for maybe the dozenth time, and have since spent time installing and configuring stuff so that I could really make the switch to that platform as a desktop OS.
So far so good. I love Gnome 2, and the gcc 3.2 seems like it makes much faster apps. I even pulled my good old Trillian-2-Gaim buddy list converter out of the closet to IM random buddies and spread the word.
I shall get back to blogging regularly when the dust settles down and when I fix the annoying issues with sound support (I have sound, but it makes a quack before making sounds, which I believe can damage my speakers).

I saw a guy posting the code he used to syndicate his phpBB2 forums, and I thought it would be nice to just improve it and make it a very easy mod for phpBB2.
The result is there, a file to save in your phpBB2 root and 4 lines to add to another file, and your forums are syndicated.
Ah, the warm fuzzy feeling of RSS 0.92…

You know Jesus Mosaics ? Now you could achieve the same effect with any images to make a mosaic out of any picture (doesn’t have to be Jesus, really), using commuter, a command-line utility.
(Link from the Need To Know. Also from the NTK: what would Homer Simpson say ? Doh !)

They’re crazy about RageBoy !