Hi,
Your points are actually rather good, and I'd like to work on it.
Especially since I am currently changing a lot in my libraries and this
could find it's way into a common lib. One problem might be that I tend
to change stuff as I need it, as I am my only user ;-)
Regarding iteratee/enumerator/iterIO: I am currently using enumerator
to (re-)implement my data sources. Once you wrap your head around it,
one can use them without too many problems. One problem I have is that
it is not quite clear what library would be the "best" one. I basically
chose enumerator because of better dependencies (iterIO depends on
SSL...) and many supporting packages.
Gruss,
Christian
PS: I'll be at alternative splicing, not BOSC -- it's closer to our
work.
On Thu, 09 Jun 2011 12:20:53 +0200, Ketil Malde wrote:
At least Christian and I are going to ISMB/ECCB, and the preceeding
BOSC
and 'codefest' - i.e. a bioinformatics hackathon. We should start
thinking of some tasks to address, and there's a common document for
the
codefest (see attached main from Brad Chapman).
Some things I'd be interested in:
* Separating out a bio-core library. There is a bit of overlap
between the different libraries, I think there's a market for
a
small common library with the most common, shared, data
definitions.
* benchmarking the bio library. I used to have a hacked
quickcheck
for this, but it doesn't work after moving to QC2. A better
approach is to use criterion for this.
* Iteratee-based parsing and analysis. I really need to
experiment
with iteratees and enumerators to get a better feel for how
they
work in practice. This is likely to be more a learning
experience than resulting in useful code, but who knows?
* Cleaning up parser interfaces. Unless we decide to move
everything to iteratees, I'd like all file formats in bio
to follow the '{de,en}codeFoo,{read,write}Foo'(etc) pattern.
And of course, the perennial task of improving documentation and
adding
test cases...
Also, broadening our horizons a little:
* Look into packaging of libraries and applications (rpms and
debs),
especially 'flower' and 'flowsim', but perhaps others as well.
* Integrate applications with Galaxy or other platforms.
This is just some stuff I thought of, doubtlessly I've missed many
important and interesting things, so please suggest or vote for
whatever
tasks you prefer to see worked on!
-k
_______________________________________________
Biohaskell mailing list
Biohaskell@biohaskell.org
http://malde.org/cgi-bin/mailman/listinfo/biohaskell
