patai_gergely:
I would concentrate on the fact that when you use Haskell, you write code that is less prone to errors and bugs. When you write a program in Haskell and it finally compiles, chances are that there are far less bugs than in a program written in another language
the question is how to justify this in 1 hour. technical people don't buy such arguments with justification. but if it will be done, it would be best presentation possible That's not really the direction I'd like to take anyway, because as I said, the primary goal of this talk is not to convince these people to use Haskell, but to give them an overview that's "more helpful than reading a book", and it should definitely introduce specific language features and important patterns besides the general talk. The question is which topics are noteworthy. But here's some more context for you:
My talk will be the second (there's one every week). The first one will be a comparison of Prolog and Haskell by an expert of the former (not surprisingly a rather mathematically minded fellow), and the third one will be about monads in category theory. Also, the initiator of the seminars is mostly interested in processing natural languages, so it might be a good idea to bring up at least one relevant example -- for instance, the latest post on sigfpe's blog could be nice, but it comes with too many prerequisites.
See also Simon Peyton Jones' intro to Haskell, recently: http://erlang-factory.com/upload/presentations/116/SimonPeyton-Jones-ErlangF... http://www.erlang-factory.com/conference/London2009/speakers/SimonPeytonJone... And one big difference from Prolog: centralized community resources. One main implementation, one distribution mechanism, one build system, one central library archive. -- Don