Subscribe to Computing Intelligence

Tuesday, June 30, 2009

Software Complaints: Mathematica and UTORwebmail

Yesterday was an exceedingly frustrating day at the Institute, primarily driven by one piece of software: Mathematica. I know that learning how to use a new piece of software can often be frustrating, and so I have been trying to reserve my judgment. However, there are a few things that I feel justified in complaining about now, because I know they will drive me nuts for a very long time to come (and there is very little excuse for a professional software bundle to have some of these problems). While I am complaining about software bugs and oversights in Mathematica, I thought I would throw in a complaint about the University of Toronto's webmail as well (just because it has been irking me for years now, and, as has been pointed out in the past, part of what blogs are for is complaining). For the record, I am using Mathematica 6.0 on a Linux system (KDE 3.5.7), so it is possible some of the issues might be compatibility issues between Mathematica and Linux.

To start, Mathematica seems to be ridiculously unstable. I had it crash with a segmentation fault twice yesterday on startup (in other words, I hadn't even opened a file and started doing any computations yet). To make the instability worse, Mathematica never seems to do any sort of autosaving (sure, one should not rely on autosave, but it is a courtesy). The lack of an autosave is exacerbated by an inability to 'undo' and 'redo' more than a single change (and what Mathematica registers as a change is hard to predict). All that combines to make any sort of code development far more unpleasant than it ought to be (using the term code loosely - Mathematica is not really programming, which is also frustrating, but I recognize that program development is not exactly its primary function).

A relatively minor thing that has really started to get under my skin, though, is the fact that Mathematica opens a lot of windows (if you need help on the use of a function, you can highlight the function name and hit F1 to bring up a documentation window on that function - this is actually a very handy feature). This in and of itself would be tolerable, if slightly messy, but every time you close one of those windows the program for some reason resizes one of your other windows that remains open. Other than sheer unprofessional oversight on behalf of the software developers, I cannot imagine why this should happen. Even if your only remaining window is maximized, it will kick out to some oddly shaped default window size. Although this is not that big a deal (it's a simply matter of resizing or maximizing your window again), when you are frustrated with syntax or in deep concentration over the underlying mathematical motivation of what you are trying to do, it is horribly distracting.

The rest of my gripes are probably more related to unfamiliarity with the software (for example, I find the syntax to be ugly and cumbersome, but that may change), so I will turn my rant to UTORwebmail now (in fact, this is a problem with the entire UTORid system). In all the years I have spent at the University of Toronto, I have yet to figure out why the login screen times out after two minutes "for security purposes", but the only way to logout is to close the browser. Someone got that completely backward.

Edit: Wolfram responded to my post, as I describe here.