It's been nearly a month now since the Edinburgh Hackathon (29 & 30
August), organised to coincide with the convergence of Haskellers
arriving for the ICFP. My apologies for this very late summary!
Eric Kow was the driving force behind this Hackathon, though I agreed
to help out locally because I live in Edinburgh. We had trouble
organising a good venue through a combination of university holidays
and Edinburgh International Festival. We spent the first day camping
out in a cafe. On the second day Graeme Hutton arranged for us to use
an empty room in the bowels of the ICFP venue.
If I remember well the first day was heavily Darcs-oriented, and there
were interesting conversations on version control representation,
conflicts, and the future of Darcs over breakfast. The staff at
Henderson's Restaurant didn't seem to mind our impromptu geeky
enclave, and we kept buying coffee to pay for our wifi usage.
I'll let the other Haskellers explain in their own words what they
were up to for the rest of the time. Maybe some of them have updates
on where these projects have gone since then?
"At the hackathon I started working on a small extension to the GHCi
debugger. The goal of the extension is to allow the debugger to show
the term representation of thunks where possible. Presently, to find
out the value of a thunk in the debugger you must force its
evaluation. However, this is not always desirable, especially if the
thunk contains breakpoints, or if it diverges. In some cases it ought
to be possible to extract source code information from a thunk and
show that to the user. I didn't get very far in the implementation
because I spent most of the time re-learning how the term printer in
the debugger works (and chatting to other people)."
-- Bernie Pope
"I was continuing to work on my camp correctness proofs. In
particular, investigating how best to handle symmetric lemmas, and
refactoring the existing definitions and proofs."
-- Ian Lynagh
"I worked on fixing Solaris/SPARC port of GHC. It turned out to be a
tedious header file related issue, not very interesting."
-- Ben Lippmeier
"I worked on the formlets (again), I released a version that doesn't
add any functionality, it only solves a bug."
-- Chris Eidhof
"I worked on a feature to let darcs users more conveniently upgrade to
the hashed repository format. I also had lots of interesting
discussions with darcs users and hackers :-)"
-- Eric Kow
"I worked on analysing some darcs performance problems, fixing a
couple of GHC bugs, and some discussion about packages."
-- Simon Marlow
"I (along with Benedikt Huber) worked on adding a 'cabal init' command
which automates much of the work of cabalising a project (creating a
.cabal file, a Setup.hs file, a LICENSE file, etc). It's not quite
done yet but should be finding its way into the cabal-install HEAD
soon!
-- Brent Yorgey
"I worked on the new hackage-server a bit.
darcs get http://code.haskell.org/hackage-server/
live instance at http://sparky.haskell.org:8080/"
-- Duncan Coutts
Thanks everyone who took part and who supplied an elegant summary of
their work that weekend. I hope you all had a good time!
Cheers,
D
--
Dougal Stanton
dougal(a)dougalstanton.net // http://www.dougalstanton.net
