|← Fun in Herts, May 2008||Fun in Leeds, March 2009 →|
Olaf and Simon said
Have a break and join us for some fun on Monday, 17 November 2008.
Below you find the full programme for the afternoon, followed by a description of how to get here.
Please send me an email (O.Chitil@kent.ac.uk) stating that you come and whether you will also be here for lunch and/or dinner, so that I have an idea of numbers (where several people come from the same institution, a single email giving names would be appreciated). If you come by car, please contact Jenny Oatley for a parking permit (see below).
We are looking forward to welcome you in Canterbury,
Olaf and Simon
Hello fun people,
term time has or is starting for most of us and we are looking forward to a fun break in the middle.
On the afternoon of Monday, 17 November 2008, Fun in the Afternoon will take place in the Computing Laboratory on the Canterbury campus of the University of Kent. You can get a first idea of the location here, but details of how to get to the venue will be sent later.
Now please think whether there is something you would like to speak about. To offer a talk for Fun in the Afternoon send me a title and an abstract of about 150 words by Monday, 20 October. I'll let everyone know the programme of selected talks soon afterwards.
Some people expressed an interest in staying over-night in Canterbury. We'd certainly be happy for informal chats on Tuesday morning and Canterbury also has a lot of tourist sights to offer. Please book your own accommodation. The following is a list of B&B's and small hotels that we frequently use for visitors; they are all located on the axis between town centre and campus and thus easily reachable on foot:
How to get there
Maps to find the Canterbury campus are located here.
On the map the Computing Laboratory is J3 and the Gulbenkian theatre (where the Gulbenkian cafe is) is J10, right next to each other.
If you come by car, please email Jenny Oatley (J.A.Oatley@kent.ac.uk) with your address to receive a visitor's parking permit.
Train: Note that most trains from London to Canterbury divide shortly before Canterbury, with either back or front half going to Canterbury. This is announced numerous times in the train.
- Trains from Charing X / London Bridge arrive at Canterbury West station. There you can take a taxi or walk 20-25 minutes to campus. To walk, just turn left when leaving the station and follow that road. Go straight across a roundabout and 10m beyond a cycle and footpath goes left. It goes through a tunnel under the railway line. Now you just need to follow the signs to the university. It is all either dedicated foot path or quiet residential street. Note that the first bit of this path, the tunnel under the railway line, is not shown on most online maps (such as googlemap). However, you can see it on this map. Note that both the pub and the curry place of the evening are just 5 minutes walk from Canterbury West station, so it is the more convenient station for going back.
- B) Trains from Victoria arrive at Canterbury East station. There you can take a taxi or walk (10 min, following signs along the ancient city wall) to the bus station. Busses to university (Unibus) depart every 10 min, taking about 15 min.
|12.00 - 13.30||Lunch in the Gulbenkian Cafe Bar|
|Talks in the Brian Spratts Room (SW101) in the Computing Laboratory|
|13.30 - 14.15||Enumerating the rationals, revisited: an exercise in reasoning about co-data|
|Ralf Hinze, University of Oxford|
|14.15 - 14.45||Using OTP, when a language is not enough?|
|Aniko Nagyne Vig, Erlang Consulting|
|14.45 - 15.15||1st break|
|15.15 - 15.45||Fun with Agda|
|Nils Anders Danielsson, University of Nottingham|
|15.45 - 16.15||(Mc)Bride and prejudice (by a student of J. Austin)|
|Nimish Shah, University of York|
|16.15 - 16.30||1st break|
|16.30 - 17.00||Generics in Small Doses|
|Adam Sampson, University of Kent|
|Adam's slides and a link to his paper|
|17.00 - 17.15||A Demonstration of Wrangler, the Erlang Refactorer|
|Simon Thompson, University of Kent|
|17.15 - 18.15||Drink at Ye Olde Beverly|
|18.15 - ??.??||Curry at The Ancient Raj|
Enumerating the rationals, revisited: an exercise in reasoning about co-data
Ralf Hinze, University of Oxford
Around 1858, Moriz Stern discovered a beautiful way of enumerating the positive rationals without duplication. On the 27th May 1976, Edsger Dijkstra fooled with figures discovering a function enjoying several intriguing properties he called fusc. We show how to relate the two discoveries using the theory of co-data.
Using OTP, when a language is not enough?
Aniko Nagyne Vig, Erlang Consulting
Erlang is a great FP language, but when large teams develop distributed, fault tolerant, highly concurrent soft real time systems with requirements of high availability and scalability, a language on itsown, no matter how powerful, is not enough. With a powerful language, you need powerful middle-ware, especially when time to market is crucial. This talk introduces the need and advantages of using middleware in the development of industrial grade systems. It looks at Erlang's evolution, and how it helped form OTP. It will give an overview of Erlang, the design principles and components and the extensive set of libraries and tool sets that make up OTP. Behaviors are derived from the fact that processes follow a set of generic patterns. Even if they execute different tasks, they will execute them following a similar pattern. Erlang behaviors consist of the extraction of these patterns into generic modules, providing us with a well tested base to implement client servers, finite state machines, supervisors and event handlers. Some of the generic applications which today are distributed as part of OTP include database applications, O&M applications, Interface Applications, a Corba ORB together with necessary development and runtime tools. We will provide an overview of these tools, describing how they reduce development costs while increasing quality.
Fun with Agda
Nils Anders Danielsson, University of Nottingham
I will give an introduction to the dependently typed functional language Agda and its interactive development environment. Hopefully the talk will give an impression of the kind of fun one can have when programming with dependent types.
(Mc)Bride and prejudice (by a student of J. Austin)
Nimish Shah, University of York
At the "Fun in the Afternoon", (York, 2007), Dr Conor McBride introduced the idea that lists are the `new' pre-order in dependent type theory. This talk simplifies the idea by examining how mathematicians define the natural numbers in Zermelo-Fraenkel's (with the Axiom of Choice) set theory (ZFC). Points out that the construction of a list representing the natural numbers is at odds with ZFC, hence replacing a list with a semi-group/monoid definition allows natural pre-orders together with all the theorems of semigroups/monoids (carried out in a combinatorial manner) now being applicable.
Generics in Small Doses
Adam Sampson, University of Kent
Tock is a new compiler for concurrent imperative programming languages, designed using nanopass techniques. Nanopass compilers transform a program from source code to the target form through the application of a series of transformation passes. Because these passes are usually small and self-contained, the resulting compiler is highly modular, and easy to test and extend. We describe the generic programming interface that we have designed for building nanopass compilers in Haskell, and show how it can be implemented efficiently by combining techniques from the “Scrap Your Boilerplate” and Uniplate generic programming systems. The resulting generics system supports operations upon multiple types without recourse to runtime type introspection.