?

Log in

No account? Create an account
 
 
01 September 2014 @ 06:42 pm
Formal Methods for Coordinating Multi-Agent Systems  
A couple of weeks ago I once again enjoyed the "sensible arrangements" (seemingly limitless alcohol on an honesty system) at Schloss Dagstuhl. It occurs to me to marvel, slightly, at the way the Dagstuhl staff ruthlessly organise us. Seminar leads endlessly entreat us not to be late for lunch, supper, morning coffee or afternoon cake and there is the ritual of writing your abstract by hand into the book of abstracts.

This was a particularly successful Dagstuhl from my point of view and I came away with an invitation to give a talk, a revelation about a paper two of us were currently working on, discussions that may lead to two further papers, not to mention a list of references to check out that may be useful for my work.

The seminar itself was slightly derailed when a logician* announced that he couldn't possibly discourse sensibly on the topic without a formal definition of coordination. A working group was set up to come up with such a definition with... moderate success? Accounts suggest the working group spent a lot of time not talking to each other or, at least, all working individually on their own and then having 5 minutes intense argument at the end of the session. I rather liked one of the definitions they came up with but apparently no one else did.

There were three working groups in total but the other two quickly combined, together with some refugees from the "let's define coordination" working group. We decided to tackle the rather easier question of how you should go about engineering a solution to a coordination problem. We were aided in this by the discovery that our meeting room had an entire wall that functioned as a white board** with which we (well I, because I appropriated the whiteboard pens) had a lot of fun.



Our whitewall. We did a lot of recording of the white wall since we have vague plans to write a position paper based on it.


A Dagstuhl Panorama - because.



*I was tempted to write a Russian logician since most good logicians in Computer Science are Russian. However this particular logician was Bulgarian with a Ukranian surname (and a family history involving the White Army) who was about to move from Denmark to Sweden. So I'll just go with "a logician".

**This probably also counts as a sensible arrangement.

This entry was originally posted at http://purplecat.dreamwidth.org/125923.html.
 
 
 
Mayrain_sleet_snow on September 1st, 2014 10:12 pm (UTC)
I once handled front-of-house for a very small conference and find I have every sympathy with the Dagstuhl staff. It's like herding cats who function entirely on caffeine.

Also, sounds like a most productive and rather fun time was had. :)
louisedennislouisedennis on September 3rd, 2014 09:55 am (UTC)
It was a particularly good Dagstuhl from my point of view. I was very glad I went.
sophievdennis on September 5th, 2014 08:59 pm (UTC)
If you want your very own wall-wide whiteboard, check out IdeaPaint
louisedennislouisedennis on September 5th, 2014 09:04 pm (UTC)
When we redecorate my office and assuming I can actually figure out a wall that doesn't need to be covered with bookshelves, then it is definitely going to be whitewall.

Although I'm tempted by a white office, tbh, I'm sure one can always use the odd corner for scribbling in.