Archive for 2008

The Gengo-plugin for wordpress claims to support multilingual blogging, so I’ll give it a try with this utterly contentless post (which exists in english and german). Let’s see what happens!

5 Minutes later… There is no way to switch the current language in any human-compliant way. Gengo does come with two widgets, one displaying just a static list of available languages (wow.) and the other is supposed to (judging by the name) switch the displayed language, but due to major bugs in the JavaScript, it never works. And of course there is no fall-back to something non-scripty.

So I’ll keep up with my way of tagging the posts with their language code and some not very sophisticated magic in my custom theme to display some German messages, where appropriate.

On the non-fail side: another plugin/widget puts my currently listened-to albums somewhere (with JavaScript look at the top edge of the screen, without look at the bottom of the page). Yes, it’s really useless but I swear there won’t be anything like the mood-descriptions or lolcats. Ever.

Google has the annoying habit of ignoring the language-setting in your browser and instead locating its users by georesolving their IP-addresses. Thus I get this message on accessing gmail.com:

If you’re traveling in Germany, you can access your mail at http://mail.google.com.

Oh, and we’d like to link the URL above, but we’re not allowed to do that either. Bummer.

I doubt this text will stay online long :)

People might call it unprofessional…

I see no use at all in Snap, the annoying service that pops up a postage-stamp sized, way too high compressed, and thus completely useless screenshot of the linked-to websites when the mouse resides on a link. It only does so on websites which include Snap’s script, when this would be so much better suited as a browser extension that can either be disabled (or just not be installed) or that works on all websites.

This annoying website of mine, probably shows a pop-up message when viewed with Firefox under Windows, whining about missing fonts nobody needs, cares for, nor is in any way likely to download, just because some suspicious-looking pop-up said so. And I’m tremendously pleased to announce, that it also does so on Snap’s screenshot-cluster, which then looks like this:

Snap screenshot of this website, with a whining message on top of it.

Spectacular. I wonder, though, why the window has no title-bar… They seem to use Windows, or some Redmond-themed linux without a window-manager…

It’s really complicated! Like, when you don’t send the first letter, it totally doesn’t work…

$ curl -I http://abc.net.au/gruentransfer
TTP/1.0 301 Found
Location: http://www.abc.net.au/gruentransfer

Today I found (via bunny) Eliza Frye, who has a nice style (and seemingly one of the few web-comic artists not working completely digital? But I know too few comics to say so in earnest). So nice in fact, that I had to copy one of her motives using Inkscape, to check if I’m still any good at vectorizing:


(If you see nothing here, get a browser that does SVG!

See the original, or my version in big (scale your browser window!), gradient added for sheer fun.

She released a wallpaper in higher resolutions — nice.

Thanks to FUSE it’s unbelievably easy to implement new and awesome filesystems for Linux, such as for example MetaweblogFS, which allows me to write this post using cat > $title && touch $title (hopefully).

Ok, so the touch part did not work, but at least it posted a draft. Tags were missing, anyway. Looks like I won’t be a user of this particular filesystem in the future, but it was fun to try.

This seems like great fun, SSHfs looks even remotely useful and maybe TagsFS, which presents your mp3 and ogg files in a virtual tree based on the tags of the files might even solve my music-library woes.

Be prepared for xml-flatfileFS, implemented in OCaml!

es.war.nicht.alles.schlecht.the-next-great-leader.de ist ein weiteres glorreiches, mondänes, ökostalinistisches Terrorprojekt der Zenzizenzizenzic Armee Fraktion.

Dies ist Teil der viralen Marketingkampagne einer ernsthaft geplanten Sendung:

Das ZDF hat sich die Rechte an einer Castingshow für Nachwuchspolitiker gesichert. In “The Next Great Leader” will der Mainzer Sender “neue, unverbrauchte politische Hoffnungsträger” suchen. [...]

After having read a rather evil article about their newest album, which essentially praised it for being their first that’s not utterly boring, I listened to some excerpts on last.fm and immediately bought it, less than 30 minutes after reading about the band for the first time. (To my defence: I discovered two Portishead-remixes on my disk later that day, I simply forgot about them somehow as Rhythmbox forgets about my ratings after some of its frequent crashes)

To my astonishment, I received the CD (yes, these physical objects. Not trusting my non-existing backups I don’t purchase virtual goods) from Amazon on the very next day, proofing yet again that paying more for Prime or 24h delivery is a useless waste of money.

Read thorough reviews elsewhere, I’m not good at this, anyway, and as it’s their only album I know, there’s nothing to say for me about any development etc.

I like
the deep sound, speaking of both ambience and audio spectrum; the singers voice seems quite airy and feeble, which is a nice contrast to the more direct music (exactly the reason why I love Metric, I’ll write about that later, maybe)
I think it’s odd
though, that some of the tracks end in such an abrupt way that I suspected the ripped (FLAC, of course) files to be faulty. They weren’t. Also some tracks begin or end with way too long passages of pure sine waves at unbearable frequencies, bad for the beheadphoned ears.
Conclusion:
Very nice; worth the money (although the CD has no added value at all except for being a backup medium for the rips, the booklet’s empty except for a photo of a radio station and credits); best suited for night-time background listening while working (as it’s a bit too minimal if you do nothing but listen…)

I used to think letting the browser rescale embedded images should be a punishable offence, but it turns out there is just no other way if I want the content to have a width relative to the text. Thus a 1024px wide image, scaled to (on my screen) 650px (see I planned for higher resolution!). Hit me :D

After my most miraculous recent success of embedding MathML despite Wordpress’ continued effort to destroy anything I feed into it, I got a little carried away and tried it with SVG:

Yeah, XHTML! 1 + x + x

As the Firefox-Developers re-enabled foreignObject in the current builds, above image uses the technique to embed MathML, which in turn embeds XHTML (the link) and another SVG image (the hand-drawn text). And it works! — at least it does in Firefox 3, but who cares about the others anyway?

…sehen wir doch mal nach, ob MathML mittlerweile wieder funktioniert:

- x ϕ t t = Φ x


Ergebnis: Gar nichts? Liegt es am Content-Type: text/html? Denn andere Dokumente werden angezeigt.

Mysteriös! Damit ist aber schonmal keine Langeweile für die nächsten Minuten (Stunden? Tage?) angesagt. Immerhin.


Merke: Content-Type matters!. Falls es jemanden interessieren sollte: In der Wordpress-Templatedatei header.php einfach folgenden Code am Anfang der Datei einfügen:

<?php
  header( 'Vary: Accept' );
  if( stristr( $_SERVER["HTTP_ACCEPT"], "application/xhtml+xml" ) )
    header('Content-Type: application/xhtml+xml; charset=UTF-8' );
?>

Natürlich lässt sich das noch erweitern, sodass die richtigen Doctypes geliefert werden, um die Browser in den Standards Compliance Mode zu schalten… Das Ganze hat auch zur Folge, dass Gecko jetzt seinen XML-Parser (statt des HTML-Tag-Soup-Ratevehikels) verwendet und sofort meckert. Mal sehen wie schlimm Wordpress den Inhalt verwurstet (die Correct incorrectly nested XHTML-Funktion habe ich schon hassen gelernt)…


Und was sagen denn die anderen ernstzunehmenden Browser eigentlich dazu?

Opera
kennt plötzlich die HTML-Entities nicht mehr, scheinbar weil sich ihr XML-Parser nicht um den Doctype schert. Nicht mein Problem, sieht ein bisschen blöd aus.
Webkit
(in meinem Fall Midori, nicht aktuell) bastelt eine Fehlermeldung in die Seite hinen, weil es offenbar ebenfalls den Doctype nicht liest, behauptet, es würde up to the first error rendern, stellt dann aber doch alles andere dar, auch normale HTML-Entities.

Stellt sich natürlich die Frage, weshalb die beiden eigentlich den Accept: application/xhtml+xml-Header senden, wenn sie nichtmal einen richtigen XML-Parser besitzen.


Oh, und wenn obige Formel in Firefox hässlich aussieht sind wahrscheinlich die eingestellten Schriftarten daran Schuld (insbesondere wenn man die offizielle Anleitung befolgt). Setzt man, z.B. per about:config die Eigenschaft font.mathfont-family auf STIXNonUnicode, STIXSize1, STIXGeneral, Cambria Math, Standard Symbols L, DejaVu Sans, wird es doch ganz nett.