Out of curiosity, Gunnar Grimnes compiled the size of the NEPOMUK codebase.

For the dfkinepomuk branch, where we mostly work on, the stats are:

• lines of code (loc): 539485 (including blank, comment, licenses)
• java files: 1725
• generated loc by rdfreactor: 202113
• largest file generated probably: 30k loc in one file

The current protests against the Iranian coup d'etat and election fraud are by people like us. Allededgly, the official results were 12% for Ahmadineschad and 62% were announced. Iran has more people under 30 than in our country, they read the web, they blog, they want to live their lives without gov threats.
I pray for them.

At the moment, the protests are peaceful and heart-touching things happen, such as these protesters protecting a captured policeman:

When you drill down far into Java reflection, you end up at the reflection class for generic types. This is, if you have

class Blah {
  T t;
  public T get() { return t; }
}
... Blah stringBlah = new Blah();

Then you can get the configured return type "String" by reflecting on the method "get". The return type of that method is String. BUT if you need the generic variable T of Blah, you end up calling Blah.class.getTypeParameters(). And here the fun starts in the doc of TypeVariable:

The type parameter D represents the type of generic declaration that declared the underlying type variable.

wtf? this is as complex as playing braid.

On Friday, 5th June 2009, I defended my PhD on "The Gnowsis Semantic Desktop approach to Personal Information Management". I defined the PIMO ontology, an architecture built on gnowsis 0.9, and evaluated it. My conclusion is: the Semantic Desktop, as I define it, supports users in filing, finding, and thinking about information.

On Friday, I defended the thesis in a 30 minute talk. I concentrated on one story-line "knowing more than you can remember" and knowledge articulation. Thomas Roth-Berghofer passed on a tip by Professor Richter: have one slide in the presentation that is really complex, to show that you did something challenging. So I drilled down on the dropbox application to show the innards of the system. There was a misunderstanding between one Professor and the School of Informatics about the date, so we had to wait a bit until he finally arrived, but luckily everything went excellent. After the talk the Professors debated about the grade and then called me in, this raises the anxiousness effectively. They decided to grade me "sehr gut", which translates to "magna cum laude" and is the second-best grade (after "summa cum laude").

For preparation, I collected the major arguments that needed to be in the presentation, cut away all the details, took a lot of my slides from my previous 85 talks about the topics, and changed everything to give one coherent story with coherent examples. I also used a few structuring tricks, such as "in-between" slides to separate areas and nice rounded corners. Here is the presentation on slideshare:

Our tradition here goes on to meet at the institute, drink some sparkling wine and give cheers to the candidate. Professor Andreas Dengel, my supervisor, gave a very nice speech about my work and my personality. I gave thanks to my peers, God and Jesus, and to Ingrid, my wife. Then the tradition is to give the candidate a doctoral hat that is built by his peers. Here is the moment:



My hat is awesome, it has a burning man in the middle, is made of tape sculpture, and glows in blacklight:

click the picture to read the details.

We then ate good food from the catering company "Klein-Partyservice" who are locals from Kaiserslautern. I also brought three crates of beer, which was more than enough for the 30 guests. Part of the celebration was opening a bottle of Barolo wine from my best friend Ebo, which tasted excellent. In the night, a few of us went to the frohlocker.de party in Kramladen and we had a lot of fun with the crew there, like bringing some good wine. And the robot:




Max and Heiko slept in Kaiserslautern and we all had breakfast together at my place, great:


This is the last important step of my three-year enterprise to blog about doing a dissertation on Semantic Desktop at DFKI (on that page you find trackback links to all steps).
The finishing step will be publication as book and then receiving the title.

Today I will defend my Phd, which is another point in the long story I blogged about doing a phd.

Things to do 2h before going to the defence:

• test your talk again. do it before good friends who you trust and who will give you positive feedback. In my case: Thomas Roth Berghofer and Olaf Grebner
• but 3 crates of beer and 12 bottels of sparkling wine into the office refridgerator
• use a lot of axe deo-spray to fight cold-sweat-of-panic
• print your presentation slides in case armageddon happens and no beamer is available and you have to give your talk without beamer (thx to Olaf for the tip)
• blog

Then, at 4pm, go and defend your thesis. To put it in starcraft-speak: The attackers will watch you build your base while they have enough time to go for resources and then do a zergling rush. So put your arguments into bunkers and use your tanks for cover fire. Anyway, they are not in for winning, its the joy of attacking you, remember that.

see you on the other side....