I love Bioconductor, I really do. It’s a great framework, nice set of tools that do anything that you want (not that I would want to do phylogeny with it, topic for another post). But as anything scientific it has its drawbacks, and I can point three of them:
1 – Installing packages – we’re having problems running one package, what requires us to modify R’s compilation parameters from time to time, uninstall and reinstall things. And the most annoying thing is when you have a fresh R binary and framework, you do a basic install first and when you want another package installed if it has dependencies that are already in the system, they are downloaded and installed again.
2 – Website – Looking from the design and navigation point of view, it’s good. Looking at the functional side is bad. Sometimes you’re looking for the package page right away and Googles’s first hit is the package’s page, but for an older version of R and Bioconductor. Maybe if the URL for the current release were different than the legacy ones, would be easier for the user and for indexers to get the actual relevant item. Of course, some people might be looking for legacy stuff, and if you put links for the other version in the package’s page, everything would be easier.
3 – Documentation – That’s not Bioconductor team’s fault, but some of the packages have very poor documentation.
Apart from that, I’m quite happy to have an open source powerful toolset for my data analysis. Well done.
About this entry
You’re currently reading “Bioconductor,” an entry on Blind.Scientist
- 02.10.11 / 8pm
- Bioinformatics - opinion