Thursday, September 02, 2004

UPS-y days, eh?

UPS delivered a book for me on August 31 -- but not to me. The customer representative was friendly, and figured out what went wrong.

It took me forever to figure out how to get parallel Matlab running on the local cluster (though Viral did most of the work). Eventually, I discovered it was an error in my .cshrc file. In the process of figuring that out, though, I learned some useful things about ssh, ssh-agent, and ssh-add.

A gentleman using my Matlab interface to the finite element code FEAP has been having persistent problems; I think this is because he is using an older version of the code. But he has done an excellent job in presenting his difficulties -- most people give too much information or too little -- and may go around the problem by upgrading his version of FEAP. And now I know another person interested in using my software.

A colleague is having problems getting results from ANSYS to match results from Sugar on some simulations. I have not yet determined the root of his problem, but I think I have a lead. If I'm right, it's not a bug in my software (hooray!), but a subtlety in the modeling. Modeling issues are much more interesting than software bugs. In addition, we have both learned a few things about frequency shifting due to residual stress -- a topic which I think does not relate to the mismatch between ANSYS and Sugar, but may explain some of the mismatch between the simulated and the theoretical results.

I looked up the warranty information for my laptop. The warranty period ended in May. But from the symptoms, the problem with my laptop is related to the backlight, which may have broken or may have simply become unplugged. Taking apart the screen to find out should be educational; and new backlights are cheap.

I am no Polyanna (though I could pretend to be a polyhedra if that would make a proof easier). But some days the world seems brighter for the bugs in it.