Learning and relearning
We ponder an example of wasted time in science.
Our astronomer has time off this week, much more than usual, so he has picked up a project he’s been working on for a long time. “Working on,” however, is a misleading phrase in this case, since the pile of papers has been gathering dust in a corner, and the electronic files have been sleeping on a rarely-used computer. Indeed, one afternoon had to be dedicated to recharging the battery on the old machine.
The project has to do with visual photometry of variable stars, and we may give you more details when we can get our astronomer to write them up. It requires analyzing a pile of data in several different ways, and then presenting the results in a coherent manner. There’s a lot of work involved, and it has evolved over time as ideas have come (and sometimes gone). It also relies on a certain software package that, while powerful and very useful, has a syntax quite different from any other we’re aware of. In particular, our tutor’s expertise in Java is no help.
Our astronomer has let it go so long because, once he set it down, it required a significant time to find his place in the process and to relearn the software. In the normal weekly schedule he would no sooner have worked out what he needed to do next than he’d be taken away to work on something else. It’s possible he could make some progress; he has intended to get back to it for some time; but he rather dreaded the amount of fragmented effort he’d have to put in. Now, with several days straight to work on it, he hopes to get it in shape for submission to a journal. In this case, even a little extra time can result in a lot of progress.
It occurs to us that our astronomer has been highly inefficient in his use of tools and techniques, especially software. In grad school he learned C to do some calculations, then another language to display things in three dimensions, and finally produced some results in Mathematica. One would think that three languages would be enough, but as a postdoc he picked up the standard image-processing software IRAF to deal with his observations (a program built by and for astronomers, and sometimes described as “user-surly”). For further investigations he found other software more useful. So in the end he has learned and forgotten a new language for every couple of papers. This is highly inefficient; learning and relearning can take a lot of time, time better spent actually doing astronomy.
We do know of scientists who develop a technique and stick with it. Maybe our astronomer should concentrate less on coming up with new ways to do things, and more on using the tools he already has.
