Profile cover photo
Profile photo
Christopher Fields
347 followers
347 followers
About
Christopher Fields's posts

Post has attachment
Christopher Fields commented on a post on Blogger.
I can definitely relate to that XKCD. We (bioperl) have a ton of code that's heavily API-bound; minor changes in significantly used modules cause someone a headache down the road.

My general workflow for bug checks:

1. I check the test first to see if it's sane. I've found many that were just wrong, simply parroting output. Quite a few had obvious Data::Dumper functions built-in that simply dumped what was in the data structures, which were very likely then added to test, which kind of defeats the purpose of the test.
2. I check the documentation to see if the function is actually described and behaving as documented. If it's returning the wrong thing or accepting more inputs than described, I consider it a bug and just fix it, then point to the docs if there are complaints.
3. More often than not we accept pull requests. The key exceptions: where a pull is adding a new dependency or any additional significant overhead, or when a set of modules should probably be submitted to CPAN as a separate distribution.
4. Deprecate when needed. We have a built-in hack that adds a warning to deprecated code or modules, emitting warnings at the version where deprecation kicks in and then failing in a specific version if they are called. In some cases we may simply alias a method or add a warning saying 'method x is now called method y for consistency, please use that' and simply call method y.
5. There are a fair number of experiments in bioperl that somehow made it into the code base, in some absolutely hairy ways. If I find anything that seems half-way implemented and is causing problems, I have no issues in deprecating immediately and ripping it out.

Key thing is, if its an obvious bug it makes more sense (to me at least) to address it and then ask for forgiveness later.

Post has attachment
Christopher Fields commented on a post on Blogger.
Just curious, but I don't suppose a preprint version of your RNA aligner paper would be laying around anywhere? :)

Post has attachment

Post has attachment
Very nice interactive graphic on NGS technologies!

Post has attachment
I hesitate to re-post homolog.us anymore, but this is an interesting one.

Post has shared content
Very cool data visualization work from Holly Bik (post doc in my lab) and Phinch.Org and Pitch Interactive

Post has attachment
Using that fancy 'next generation sequencing'.

Post has attachment
If you use git...

Post has attachment

Post has attachment
For those keeping up, yes Perl 6 exists, and it's looking very promising
Wait while more posts are being loaded