By that I mean this blog entry has nothing to do with the transgendered community. It’s actually about the progress of my research work at the university, or the complete lack of progress. Researching in numerical methods mean that I have to be a good mathematician and a good programmer. I can say that I’m both (okay, I exaggerated), but the person that I inherited the code from was neither.
The lack of progress was because I unearthed a few bugs in the code. Substitute Samuel L. Jackson’s voice in Snakes on a Plane: “There are motherf****** bugs in this motherf****** code!” There are serious errors in the numerical method (for those who care and understand, the problem is with applying the Riemann boundary conditions for an aerodynamic problem: the boundary is applied differently in the left-hand-side matrix from the right-hand-side vector in the linear system for an implicit time-marching method) and also with the programming (properly declaring and dimensioning arrays). Now that I’ve put them into words, my second problem doesn’t sound so bad now.
For the record, FORTRAN sucks. Anyone who must program in FORTRAN for whatever reason should get a kiss and a hug for enduring the pain and suffering. Anyone who insists on its superiority should be shot. Okay, not shot, but at least severely reprimanded. F90 helped dull the pain slightly, but the standard is 15 years old, and it’s still not fully supported by everyone. Fifteen years ago I was cheering for my new 25MHz 80386. And then the old legacy code is a bitch-and-a-half to work with. Add the combination of poor compilers and debuggers, and lack of support of the Alpha processors, life can be, well, a bitch.
To old-school FORTRAN programmers in the world, please don’t use GOTO statements anymore. I hate you.
I’m enlisting help from another lab-mate’s today. Wish me luck.
This post is also available in: Chinese (Traditional)