Recently in Weblog Category

Anil Dash has an interesting upgrade guide for Wordpress. I think this is only deserved, after the upgrade path opportunity that Wordpress used during the MT3 licence debacle.

P.S. tense twittering between Matt Wullenweg and Anil Dash. I think Anil is right here, and I cannot find anything in his original post that is wrong and even less desperate, it's a fair shot at competition.

One day a Google bot sent me a message to welcome me back to Orkut. I hadn't been there for ages and couldn't recognize the place, for all its walls were covered with some alien language I couldn't read. I then learnt that it was portuguese and the aliens were Brazilians who had taken over the place.

Fast forward six months from now, and LiveJournal's walls will be covered with cyrillic.

(And big hairy, scary bears will be wandering in its datacenter.)

My two blogs simultaneously crossed a symbolic milestone today, with just 1000 entries each! It was not planned, I just saw the count on the MT menu and realized the coincidence. In December, my first blog in English will turn 5 years old. The French one turned 4 last month. Both gathered exactly 7600 comments as of today, which gives me the warm feeling that this blog discussion thing worked out quite well (though the French are really more vocal, up to their reputation :D).

Thanks for your attention :-)

A German web security searcher has found seven XSS (Cross Site Scripting) vulnerabilities in WordPress and has launched a patch for them... in the form of an XSS blog worm!

[From Blogsecurity and Planet-Websecurity]

I'm upgrading this blog software to Movable Type 4, probably breaking a few things around :p. A few thoughts while I'm working on it (this post will evolve as I dig in)...

N.B. I'm working behind the scenes first. And this is still a beta version (4.0-rc1), but my little finger tells me that those who are waiting for the production version won't have to wait too long now ;-).

Things I'm really happy with...

• The installation assistant is a breeze (with one quirk, see below)
• Pages, using the blog templates, yeah!
• Better file manager
• Much better user management (better granularity at rights, but more importantly one can really build a multi-blogs multi-authors farm with this)
• Better templates management, code, engine (gotta love MT tags!)
  
• Much better default templates, and choice of designs and layouts out of the box
• Long rebuilds seem broken down into several short perl processes, so they don't time out on hosts that impose CPU time limits on process execution (like TxD) [strike that, my first big rebuild got killed in the middle of rebuilding Category Monthly archives (what's that???) but I was able to resume by reloading the progress pop-up], they also display a nice progress bar
  

Things that show it's still a beta...

• The drop-down menus have an annoying habit of not disappearing automatically after they lose focus
• Some template tag mysteries in the header of the Main Index
  
• I get a server error 500 when trying to display the System Information
• Sometimes an unhelpful error message pops up, saying: "an error occurred"

Things I'm really unhappy with...

• The Rich Text editor is crap errr, weird... (see below)

I wonder why...

• The Learn more about installing Perl modules link is not shown on the first screen of the Requirements Check of the installation assistant. Currently it appears only after you have successfully passed the check for required modules, which means you won't get a chance to see it if you're missing a module, since it's not even listed on the error screen (too bad!) -- Remember: $ sudo perl -MCPAN -e 'install <module>' is your friend!
  
• There is a <$MTEntryTrackbackData$> tag on the default Main Index template for each entry. Apart from doubling the weight of the home page with useless junk and making spammers life easier in collecting TB links to hammer, why on earth do we need Trackback data in there? I recommend removing that tag from this template.
  
• Some functions of the Rich Text Editor do not work in Safari (like lists, link, quote). It appears to be working better in Firefox, although...
  
• The Rich Text Editor produces tons of ugly fucking code behind the screen, especially when using Safari. So it's an old-school editor such Midas, only a slight improvement over WYSIFUC in the sense that it hides it from you (unless you select "none" in the Format menu or "HTML Mode" and scream in horror at the FUC it generates, like <span class="Apple-style-span" style="font-style: italic;">...</span> instead of <em>...</em>!). For example, on Safari it generates a DIV per paragraph, while on Firefox it just spits out <br> tags or cosmetic <i> or <b> tags (not even <br />, or <strong> or <em> tags, mind you). So, three years after my FUC rant, authoring pain has just cosmetically improved but at what cost (producing tons of cosmetic-ridden tag soup like it's 98 again)? Those who have carefully designed their blog to valid XHTML/CSS web standards will see it become invalid the second they use the Rich Text editor in MT 4, you've been warned! I'm losing hope to ever see a decent open source WYSIWYG editor for authoring web content that follows modern web standards (yes, I know that HTML4 is a standard, that's not the point at all)... Sigh! [Hold on, Laurent tells me that this behavior was present in earlier betas but he doesn't see it in 4.0-rc1, I guess I need further testing]
  

I wonder how...

• To remove or edit a link in the Rich Text editor? Looks like you have to delete the whole text with link and do it again, very annoying.
• To migrate the database from MySQL to PostgreSQL? I tried what looks the simplest path (export/import) but the MT text backup doesn't transport tags!
• To disable / prevent build of unnecessary category archives? It looks like the default install builds everything that's listed on the "First Widget Manager", including very time-consuming stuff like Date-based Category archives or Date-based Author archives (quite useless on a single-author blog like mine), although none of those are listed anywhere on the live blog and that widget is not active. In other words, where are the Archive Mapping preferences that allowed to define what archives get built in MT 3.x?
• To distinguish photos within images? In the default templates, MT4 displays as "Photos" any image file you upload, but you might upload images to illustrate posts that you don't want mixed and displayed along with other photos.
  

Six Apart just released Movable Type 4 Release Candidate. Time to start playing seriously with that beast... :-)

I haven't got time to install Movable Type 4 yet^[1] but thanks to Six Apart I've been able to access a hosted version and play a little bit with it.

My first impression is excellent, I must say that I'm impressed by the progress, no the leap, made by the MT team. MT 4 isn't a cosmetic upgrade to MT3, as we've (unfortunately) been accustomed to by 6A in the past couple of years. The impressive list of new features, the fact that the MT team has stopped diverting (therefore wasting) their efforts into two different code bases for MT and MT Enterprise, and the announcement that MT is going open source, make me think that 1) the development strategy is simpler, clearer, makes more sense therefore is more perennial than the previous one and 2) MT is no longer the ugly black duck in 6A's portfolio of products and services.

Kudos to the MT development team, and to Six Apart for reviving the potential of their “Publishing Platform”. I'm really looking forward to the production version, as well as creating some interesting blogs with it in the near future.

Note:
[1] The beta still bites a lot, and the interface is rough and unfinished. I wouldn't advise you to install it to replace you existing blog, unless you have time on your hands and are happy with truly beta quality (not as in web 2.0!) software.

Last month I found a post advising to setup Zen from spamhaus on Movable Type to filter spam. Since then, I've had an alarming number of false positives, i.e. legitimate comments that are junked on the basis that the commenter's IP is listed in one of the Zen databases. By alarming, I mean that when more than half the junked comments are in fact legitimate, it's time to junk the spam filter instead. Looks like it's not that smart anyway.

Has someone else seen the same problem?

Six Apart just released a beta version of the upcoming Movable Type version 4, and a brand new movabletype.org web site where you can download it in a couple of clicks. Be aware though that it is a real beta (not a "web 2.0 beta" joke), it has known issues and is not suitable for production use. There is a fairly extensive FAQ that goes through the new features and where 6A is going with MT (i.e. open source the core, sell "components" around it like the "Enterprise" component).

From what I've seen so far, it's the most important release for years, and I'm happy to see Six Apart show some momentum again with a product that had felt a little like the unloved black duck in the past years.

A few reports to read:
- 10 Ways Movable Type 4 Will Rock Your Blog by Jesse Gardner
- an article from Duncan Riley at TechCrunch
- Arvind offers a developer's perspective and seems very happy with the new plugins and components architectures. He also says the out-of-the-box WYSIWYG editor beats competition including his own plugin. At last, no more WYSIFUC ;-), MT 4 ships with this very much needed feature!

I'll give it a try when I can, though a look at my new schedule shows that my next "coming back from the dead" window just jumped to mid-September.

About two years ago, Bits of Freedom demonstrated that one could easily get a European ISP to pull public domain material using unsubstantiated legal threats. It only took them a Hotmail account. This week, the SNCF (Société Nationales des Chemins de Fer français, the French national railways company), thanks to Typepad and the French law, demonstrated that one can do the same with a blog post.

Xavier Moisant runs a small collective blog titled "Train train quotidien" (a French expression meaning daily routine) focused on denouncing recurrent timetable and traveling conditions problems routinely endured by the travelers on the Le Havre-Rouen-Paris train line run by the SNCF. On March 16, he posted some visuals made by a certain "Teddy" that parody several SNCF visuals, to illustrate the recurring lateness of its trains, such as this one:

Meaning, roughly, "make the trains hours late - National Company of Late Trains".

You can check the official SNCF logo and signature on SNCF.fr (it reads "Give the train advanced ideas", there's a play on words as in French retard (late) is antonymous of avance). You can also check the other images from the MS Live Search cache of the deleted post.

Pretending that the blogger had infringed their copyright on their logos and visuals, the SNCF sent an email to Typepad France, and succeeded to get him censored. Typepad promptly (on March 21) deleted the post, its images and comments, then sent the blogger a nasty email backing up the dubious claims:

Following a complaint by the SNCF, we had to delete a post on your blog "Train-train" that contained several visuals using and modifying the SNCF logo. I remind you that this is illegal and, therefore, we have no choice but to suppress this type of content." (Olivier Creiche, Six Apart Europe General Director)

Trouble is, modifying a logo when done in the clear purpose of parodying, mocking or denouncing something, rather than being illegal, seems to be backed up by several jurisprudence: Esso against Greenpeace France22 (E$$O, 2003), Arreva vs Greenpeace (2003), Danone vs Réseau Voltaire (jeboycottedanone.com, 2003), where the appeal court consistently dropped charges made for copyright infringement, reasserting the freedom of communication against commercial trademarks. IANAL, but I'd not bet a buck on the legal ground used by the SNCF in this case. French law punishes with a one-year prison and 15,000€ fine sentence whoever maliciously uses the Electronic Communications law to take some online content down.

I think it's fair to say that this blog was relatively unknown until the SNCF made this move. Now, it's on the radar of prominent bloggers, and the story and those images have been aired on several trafficky blogs. A big mistake in terms of communication, and a demonstration that the blogs dynamics is something corporations need to understand better.
A few figures to illustrate this point, using Technorati ranks of the top 5 blogs referring to the story:

• The originating blog: Train train quotidien has Rank: 122,346 (75 links from 35 blogs)
• Les Influenceurs: Rank: 688 (13,488 links from 1,627 blogs)
• Pointblog: Rank: 1,880 (2,700 links from 950 blogs)
• Versac: Rank: 4,526 (1,995 links from 554 blogs)
• This blog, padawan.info: Rank: 13,570 (817 links from 262 blogs)
• L'Observatoire Des Blogs Francophones: Rank: 18,493 (513 links from 200 blogs)

Ironically, Typepad now hosts several copies of the images that the SNCF wanted to take down. What are they going to do now? And good luck with trying to play that game with my host ;-).

On an aside note, I think Typepad France made two mistakes:
- not warning the blogger in advance, allowing him to backup the post, files and comments
- claiming that what he did was illegal (how do they know?), instead of simply saying "we're sorry but we received a complain and the law forces us to take the following measures..."

P.S. (March 24): Olivier Creiche from Six Apart left a comment here to point out his follow-up response to the blogger. I find it impressive because, not only did he apologize for the chronology of events, but he went further than the confort zone provided to web hosts by French law (i.e. he could have left things as is in a status quo) by asking the SNCF to backup its threats with a legal procedure, otherwise Typepad will republish the note in 8 days. Pretty courageous when the law clearly favors cowardliness in this case, so Kudos guys!