I’ve been getting myself into “real developer” kind of stuff lately; somewhat inspired by a welder turned developer friend and partly by the number of awesome “open development projects that are going on right now.
In astronomy the workhorse codebases for analyzing data have either been hosted at the HEASARC, or they’re developed on a mission-by-mission basis. The code is almost always developed by a small (N\< 5) number of people, one or two of which may actually be real programmers. Then it’s published and left alone. For good reason; most of the programmers are actually scientsts who’d rather be looking for giant alien megastructures rather than maintaining ancience code.
There are plenty of examples in the HEASARC code base where I’ve found an example of
FORTRAN code (yes,
FORTRAN) that I rely on during most of my analysis that hasn’t been touched since 1994. In a wide variety of cases, this is totally fine. But in some it’s getting to the point where things that are common place in modern operating systems (like having a path 30 characters long) clash with the self-documenting code (which may only be able to store paths of 24 characters). Possibly no more?
There are rumors that the HEASARC is going to be moving over to github instead of using the current version control (CVS). Which is fantastic.
There is an awesome community of people out there who are taking advantage of modern programming styles and techniques that simply weren’t possible ten years ago (#astrohackweek I’m looking at you). Things like astropy and sunpy are being actively developed at such a pace that I actually had to turn off my “Watched” flag on github because my email was getting spammed. Which is also awesome.
All of this has gotten me to dust off old code (which has been on gitHub but hasn’t been maintained at all) and really take the half an hour to update the documentation, clean up the codebase, release a new version, etc.
This has led me to the lede (which I have sufficiently buried down here).
Markdown is awesome.
I’m writing this post entirely in Markdown for the first time (I really like the developer who makes DeskPM, which is what I’m using here). I’m going to try to restrain myself from cleaning it up at all in my WordPress blog interface; we’ll see how that goes.
At the same time I’m writing this and relatively effortlessly I’m also continuing to fight with LaTeX to try to make it do what I want it to do. The difference is stark; in principle, it would be nice if we really could just write papers in some pretty interface (like Authorea) and just leave it up to someone who actually gets paid to make sure that the right number of figures show up on each page. Maybe someday…
For now, I’m going to enjoy my writing.