Biology is being held back by the low-quality code that the biologists are forced to write and maintain. Producing better programming tools means that biologists can start writing better programs.
This is already happening. The world of science woke up one day and realized that the C++ and Java their CS friends were talking about weren't helping them solve problems. So they switched to Python and Perl and now we have the human genome.
This is already happening. The world of science woke up one day and realized that the C++ and Java their CS friends were talking about weren't helping them solve problems. So they switched to Python and Perl and now we have the human genome.
Oh, that's how it happened? Weird, I thought they were using lots of C++.
Don't be such a jackass. I didn't say nobody was using C++ or Java. I was saying that many individual researchers switched from Java for things like "scripts".
Infrastructure written by programmers will use different tools than glue code written by scientists.
You don't know what you're talking about. Biology is held back by the extreme difficulty of doing informative biological experiments. If you know of biology domains in which lack of good software is the primary rate-limiting factor, please share, because I am acomputational biologist, and always on the lookout for high-impact problem domains.
BTW, the software tools for the assembly of the human genome were mostly written in C++.
I'd understood - because we are constantly told by the Folding@Home people - that the problem was getting enough CPU for simulations. Is this not actually true?
The domain of biology that Folding@Home works on, i.e. predictive in silico modeling of protein folding, is microscopic in comparison the everything else biologists study. So while they may need more computing power a lot of biologists simply need more slave labor (grad students) and grant money.
There are lots of problems in computational Biology. Some can be solved by more processing time, for some you need to have better software, and for others more experimental data is necessary.
Having spent 5 years working directly with biologists at a major research center, I can confidently say that biologists don't want better tools. They have no interest in software beyond the bare minimum it takes to confirm the result they expect to see.
True, if a bit cynical. Biologists are like any other profession - resistant to change. Many research biologists are not tech-savvy and afraid or unwilling to learn new computer-aided techniques, algorithms, tools, etc.
This is further compounded by ancient tenured PIs who went to grad school before typewriters were mainstream and a publishing and grant system that stifles innovation.
Some tools are really important; agreed. Languages are a huge fucking deal. Monstrous. I can name one startup that failed because it chose the wrong language when it started, and that "wrong language" wasn't even a bad one-- just the wrong one for the kind of work they were doing (dynamically typed, in a domain where even extremely infrequent errors are unacceptable).
This is already happening. The world of science woke up one day and realized that the C++ and Java their CS friends were talking about weren't helping them solve problems. So they switched to Python and Perl and now we have the human genome.
Programming tools are a big fucking deal.