Frank's Blog

Good Enough 

Sometimes, I get asked why---and how---I wrote a doctoral thesis. There are some simple answers for the former question: Why?

1. Stay a student.
2. Get paid to be a student.
3. If you're out of arguments, you always have the doctoral joker ;-)

The second question seems to be more complicated, but indeed is also very simple: Do your job good enough. While people often search for something to become perfect, this seldom appears to be the same to other people. Just like the 80:20 rule, focus on the important things. Here is a simple graph to depict the idea:



The horizontal axis shows the time that flies by. The vertical axis shows the amount of additional knowledge you earn each time unit. In the beginning, you quickly graps knowledge. Over time, your reach the meridian. Then, you cool down. While you're still working on the topic, something absolutely new and directly related to your work is hard to find anymore. Finally, prepare finishing your thesis. The important point is to submit your thesis just in time. While you would still be able to find additional knowledge, it's not mattering for your thesis anymore. So, that's at least how I did it. And believe me, you can apply the good enough rule to almost everything!
[ view entry ] ( 2276 views ) permalink $star_image$star_image$star_image$star_image$star_image ( 3 / 3714 )
HPI Sommerfest 

Today the HPI Sommerfest is (still) taking place. Here I have some impressions for you:


The HPI band: Poolraum Party.


Eiszeit.


Some people in front of the stage---most stayed in the background.


Revolverheld.


Even the sun appeared in the evening.

Unfortunately, I have no picture from the fourth band, Tiger HiFi, since I had to leave early. If anybody has one, please send me a copy, so that I can add it here.
[ view entry ] ( 3009 views ) permalink $star_image$star_image$star_image$star_image$star_image ( 3 / 3880 )
Welcome to the Exhibition 

I know quite some people that are a bit unsure about the things we do all day long. Instead of talking about something they don't understand (it can be found here), actually, our institute has something new to offer: An exhibition of modern art. Take a look at these pictures and objects that have been placed at nearly every possible place. And yes, these pictures/objects are all original ones...










[ view entry ] ( 2758 views ) permalink $star_image$star_image$star_image$star_image$star_image ( 3 / 3781 )
Me with a Hat 

Yesterday I defended my doctoral thesis and finally got one of those funny hats. While my thesis will appear later this month on this website, you can already find the slides here (in German).


The reviewers of my thesis (from left to right): Prof. Weske (HPI), Prof. van der Aalst (TU Eindhoven), Prof. Nestmann (TU Berlin).


The buffet.


And finally, me with a hat (in the background is Prof. Naumann, the chairman of my defense).
[ view entry ] ( 2406 views ) permalink $star_image$star_image$star_image$star_image$star_image ( 3 / 3210 )
Vienna Conferences 

Ok, while I'm waiting at the only airport I know of with free wireless, here's some wrap-up of the two conferences I've attended, BPM2006 and WS-FM2006 .

Let's start with BPM. Actually, I do like the BPM conferences a lot more than ICSOC (the other BIG conference in the domain). Maybe this has to do with the more European centered audience. This year, there have been three keynotes announced of all the big players in the market, starting with Microsoft, followed by IBM, and concluded by Oracle. Unfortunately the last one was dropped and we've heard about the BBB (Big Basic Bus) once again. So that's then 2:1 for IBM. Nevertheless, the Microsoft talk was quite interesting. It was given by Dave Green, the chief architect of the Windows Workflow Foundation. He didn't talked much about the architecture but showcased the tool support, which is quite nice if you like the vendor lock. That's because Windows Workflow Foundation builds atop of a bunch of M$ products like Windows Vista, SQL Server, BizTalk, Visio, Visual Studio, Office, the new FrontPage (can't remember the current name) etc. The most interesting part is the seamless integration of office for "manual" tasks. Let's see what the future will bring for M$, I'm leaning back in my Apple chair and watch it. The IBM talk was given by Don Ferguson, one of the 50 IBM fellows and a chief architect. The talk was just about the things you would suppose IBM is talking about.

Some interesting talks and papers I'd like to highlight. First of all the current research of the Reisig group in the project Tools4BPEL. This project already got some tools running like BPEL2oWFN and Fiona (Why do I like this name? Isn't it from the movie Shrek?). Fiona is able to check an oWFN for controllability and generate an operating guideline. If I've got some time I will definitely take a closer look. Another interesting talk was given by Skip Ellis, one of the first workflow pioneers. He talked about something beyond workflow mining that only a few people accepted to be beyond Nevertheless, his presentation style and "aura" was impressive. On Wednesday Wil van der Aalst in person gave a short paper presentation about "Verifying Workflow nets with Cancellation Regions and OR-Joins". Also quite nice, but still I'm not convinced that this is a final solution. Wednesdays talks ended early so that there was time to visit the "Hofburg". After visiting the rooms of Sissi and listening to what she did all day long (hair combing 2 hours, riding 6 hours, gymnastic 2 hours, preparing a evening dress up to 8 hours, changing her dress at least three times a day) I'm wondering what the Sissi movies are all about? The talks continued at Thursday. I can't help but I found my demo session the most demo-like one. That's because most presenters showed 12 out of 15 minutes slides introducing their tool. Ok, I had the advantage of having introduced my background as a full paper already on Tuesday. By the way, did I already mentioned my demo poster? To conclude, while the atmosphere at the BPM conference was quite nice, the talks at the WS-FM workshop have been by far more interesting.

The WS-FM workshop started on Friday and the audience changed drastically. Nearly 50% of the attendees came from Italy, another 20% from Spain, only two from Germany (including me), and one from the Netherlands (Wil van der Aalst). Wil gave a keynote about DecSerFlow. While the talk was nice and also the integration into YAWL, it was criticed a bit by the theoreticians. They wondered why only linear temporal logic (LTL) has been used. Interestingly, Marlon Dumas presented a paper of one of our students, Gero Decker at this workshop. Gero was recently in Brisbane, Australia to write his master thesis and convince the people at SAP research and QUT to utilize the pi-calculus as a formal foundation for their new choreography language "Let's Dance". And the interesting point is that Marlon managed to give the presentation without showing any equation. WOW again for the that. The most interesting talk has been "SCC: A service centered calculus" by Roberto Bruni (SCC is an homage to CCS). They introduced a nice approach to build a formal foundation for SOA's. Unfortunately, this is just another proprietary framework. He even stated on his slides that they could have been used pure pi-calculus instead, but that it would not directly express the things they wanted to consider. Ok, maybe a framework ten times as complex as the pi-calculus might help. The second interesting talk was about "Towards a Unifying Theory for Web Service Compositions". They utilized web pi-calculus and extended it with work units for transactional behavior. Unfortunately, they dropped the timings contained in web pi. I missed the talk of Schahram Dustar about "Service QoS Composition" because I had to catch my plane. I also met Christian Stefansen at the workshop, a PhD student working on a quite related topic to my own. Maybe we will have some more discussions in the future.

So, now I have to board my plane and stop writing the longest entry I've ever created. Yes, there are also some pictures available.
[ view entry ] ( 2595 views ) permalink $star_image$star_image$star_image$star_image$star_image ( 3 / 3604 )

<<First <Back | 1 | 2 | 3 | 4 | 5 | 6 | Next> Last>>