I was at Matscience a good fifteen minutes in advance. All and
sundry enquired if I had substituted Priti Shankar (my colleague
who could not make it to the conference after having
registered). I (patiently) clarified that IISc should be sending
R. Balasubramanian, as acting director of IMSc, made the opening remarks; apparently he said something to the effect that no amount of direct contact can replace the Internet, which is of course a completely true statement!
Anil Seth set the ball rolling with an Introduction to Finite Model Theory. He introduced the audience to some proof techniques in finite model theory, after remarking how finite model theory differs from its infinite counterpart.
Neil Immerman followed by a talk on Logics that Capture Complexity Classes. This work is really more than a decade old and it was disappointing to see some one of Immerman's stature spending his energy on elements of computational complexity. Of course, people who had not seen this kind of stuff before should have benefitted. But the organizers could/should have allowed Immerman to choose an advanced topic.
The afternoon started with Anuj Dawar speaking on Finite Variable Logics. He spoke on the importance of being stingy about variables and how they should be treated as resource.
We dispersed for tea and there, Anil convinced me to be the chair for the last talk. I reluctantly agreed and practiced the name of the speaker for a couple of minutes with the speaker's help.
The speaker, Martin Grohe, gave an astounding talk. He spoke on
(without getting technical) treewidth, games, planarity and
showed how to relate properties on planar graphs with restricted
logics. We finished the day on a highly satisfactory note.
\paragraph{Wednesday 16th Dec}
By the next day, the drivers at Koyambedu knew who I was, my
destination etc.
Wolfgang Thomas started the second day's proceedings. He started with an introduction showing connections between logics and finte-state machines. He showed several instances of connections between logics on one side and ``machines'' on the other. He then spent the rest of the time on picture languages. It was a carefully planned and well executed talk.
Moshe Vardi spoke on the connections between FMT and alternating finite automata. Basically, he argued how looking at more powerful computation models like alternating finite automata helps in simplifying the construction of models. Moshe is obviously a gifted speaker; he made everything look very simple.
Victor Vianu brought up the last talk of the workshop. It was also a post lunch session: I will not repeat the remarks I have already made! Victor did a good job of convincing many of us that we are really doing database theory and that is where we should look to get new problems to solve.
The workshop wrapped up with a panel discussion with Anuj as a moderator. Each speaker was given a few minutes to speak. Thomas talked about the need to study specific models and special structures, and also finitely presented model theory. Vianu spoke on databases being a source of inspiration and also about viewing the web as an infinite graph. Vardi spoke of the success and failure of FMT as an area. In particular, he spoke of the growing opaqueness of the area and weak performance in the job market etc.
Overall, the workshop was of exceptionally good quality. Anil Seth and Anuj Dawar need to be congratulated for the excellent choice of topics and speakers.
That evening, I had to pay Rs 75 on my way back.