MediaWiki pushes welcome new concepts

MediaWiki is the code behind Wikipedia and despite the fact that I am not the biggest fan of its MySQL/PHP underpinnings I do like the concepts that are behind it.

First and foremost, MediaWiki does not support the concept implicitly linking all CamelCase words that are introduced into the text. In contrast, all linking is explicit which is a very welcome thing as it prevents overlinked texts that provide no contextual meaning to the reader. You might argue that a text might be underlinked then, but as it is a Wiki still, it attracts changes as long as the content is considered „not complete“ by the readers. So missing links will go away as missing information (and typos) do.

Another interesting concept is the introduction of namespaces that makes „relative“ pages possible and allows for some interesting solutions for „internal“ pages, automatically generated pages and so on without poisoning the default namespace. Pictures are managed in their own namespace and each picture is treated like a separate page but can be included by reference in other pages. This makes maintenance easy and concise.

The upcoming version 1.3 of MediaWiki is about to introduce RSS feeds and XHTML/CSS based skins so more good news is ahead. I would be happy though if somebody would port the system to PostgresQL but I don‘t see this happen anytime soon.

Installation of MediaWiki is a bit confusing however. Once you know your way with MySQL you can concentrate on setting up the Wiki itself. The INSTALL document talks about an „experimental web-based“ installation which in fact is the only viable solution. It doesn‘t offer all the options, but you can dive into the LocalSettings file and tweak it to your needs. Just don‘t try the „classic“ way described in the INSTALL files: the developers admit that it is just broken and destroys you initial setup right away. Strange.

50 Specific Ways Dylan is Easier to Use Than C++

I would have no success hiding my ongoing admiration for the Dylan Programming Language. It‘s just so fucking powerful, flexible and sexy that once you‘ve read either the Dylan Reference Manual you‘ll wonder why everybody would just think about using something else.

The Dylan Programming book is also a very helpful introduction, as it describes Dylan‘s langauge features to C, C++ and Java as you move along. Yesterday, Chris Double has summarized 50 Specific Ways Dylan is Easier to Use Than C++ in response to Scott Myer‘s book Effective C++: 50 Specific Ways to Improve Your Programs and Design. The comparison is a blood bath for C++. What was to be expected as Dylan was invented to free the world from C++ (Andrew Shalit).

If you are interested in Dylan you might want to check out the Gwydion Dylan project home page, Dylan Wiki or the #dylan IRC channel on freenode.net.

Sender Policy Framework

The Sender Policy Framework seems to be the solution we have all waited for for a long time. By simply adding a single DNS record (type TXT) to your domains, you can announce which computers are allowed to deliver e-mail from your domains. So it is checked by the receiving SMTP server before even accepting the content (also significantly reducing bandwith usage). While this is probably not the end to Spam in total, it might be the end to Spam as we know it today: coming in masses, forging your e-mail addresses, difficult to trace back, severely annoying.

The SPF guys have some excellent documentation material on their web site. There are some insightful slide shows as well: a detailed introduction teaching you the different scenarios, a detailed explanation of the anatomy of a SPF DNS record, a hacker‘s intro to What is spam? („spammers feared to develop weapons of mass destruction!!!) and more. In addition, you find a setup wizard makes setting up the DNS TXT record easy. The SPF Adoption Roll tracks popular domains for their SPF support. So far a bit more than 10000 domains world-wide have adopted SPF – but among them you find important major mail providers like AOL, Hotmail and others.

To round up the good documentation there is the SPF Wiki. Let‘s all support SPF today! It can make a difference.

April Fools Suck

Yawn. The first of April has become the most boring day on the web, beating Christmas by a wide margin. And no, I am not going to present a list of the silliest stuff. I will make sure I won‘t to switch on my computer on at all next year.

I am not alone.

Some other possible solutions to the problem:

  • Bring out AFML (April Fools Markup Language). Making it an XML Namespace, it might help introducing April Fool context to each news item. Possible elements:
    • JOKE: This entry is a joke and can safely read or ignored.
    • BAD_JOKE: This entry is a rather rude joke and might harm or confuse readers or drive them into installing viruses.
    • WISH: This entry is not true but somehow expresses something I always wanted to have but so far nobody was able to provide.
  • News reader support. Add a „Mark all April 1st news items as unread“ or „Hide all April 1st news item“ feature to news reader programs
  • Your idea here. Please comment.

Sigh.

Projekt Jäger 90

Okay. Here we go again. I just registered myself for the next Berlin Marathon in September. This time I want to be significantly faster than last year. To be specific, I want to make the 42,195 km in less than 90 Minutes. So I call this Projekt Jäger 90. Let‘s see how it works out.

Maybe I will repeat the Real Time Tracking Project I did last year. But I‘ll need much better gear to do that as last time the batteries lasted only half the time and my backup GPS system lost track in the third quarter of the route.