Wednesday, November 04, 2009

Dear Miss Manners

Q: X, at the pain of sounding demanding, is this correct - i bid 225 points for friday am ola and still got thursday am!? :(

A: hi Y,Z also has a similar problem. But I think we can work it out - OLA was one of the subjects which wasn't part of the required class demand survey the ARC and STRAW released a couple of weeks back. With a class demand survey kasi, we'll be able to find out before hand how many slots we should allocate per section. For OLA, since this didn't happen, Prof. A just estimated that there'll be about 115 students taking OLA (and this 115 will be divided by the 10 teams). Pero after the run was conducted, nalaman namin na about 7 or 8 students and hindi nabigyan ng slots. This means that the admin underestimated the demand for OLA. Prof. A was ready to make arrangements for the 7 or 8 students when we realized this problem earlier tonight. But I suggest that you send him an email na rin at [email] just so he knows. But I can help talk to Prof. A about this during reg. Pakipaaalala lang sa akin.thanks,- X

But, of course, it wasn't because they underestimated the demand. It was because the program they were using had so many bugs in it, you could have asked your local pest exterminator to clean it up.

The 225 points that Y bid above turned out to be the second-highest bid for the section. Y, along with the person who bid the highest bid (at 238) were "waterfalled" to another section, while the person with the third-highest bid (201) did not get a slot at all. Meanwhile, people who bid 90 and 60 points got the slots. So, how the fuck did this happen?

Presumably, in their desire to "improve" the registration process, and perhaps the performance of the college in general, they decided to implement new rules during registration (subject, of course, to new rules which were later implemented as they went along their merry way). Some of these are: that "crossovers" between day and evening sections will no longer be allowed and, that day students have priority over evening students on day subjects and vice versa.

When every single evening student who bid for the same subject got squat regardless of the number of points bid, you don't need to look at the source code to see that the system fucked up. Electives and OLA sections were all classified as "day" subjects, in effect, putting priority on the bids of day students above those of evening students. The result, as with all programs running with bugs, is garbage.

Given the time constraints and the indefinite requirements given them, I would say that the programmers did a relatively good job. I am assuming that undergraduate students wrote the code. To properly put all this into perspective, consider that undergraduate "machine problems" are usually given one month before the deadline. The requirements should already be clear and well-defined at this time: these requirements are what the programmers will test against. Students, of course, usually start working on these only about two weeks before the deadline. In an ideal situation, the code is tested against all possible scenarios (i.e. a stress test) presented by the requirements before it goes into production. (There is a reason why computer professionals use the word "user" when they mean "idiot".) If you don't, and the program encounters a scenario that it doesn't recognize at run-time, then you are really fucked up. Remember that for software, repair costs increase if they leak downstream - they increase tenfold with each lifecycle phase.

Computer programs are predictable. They will do what you ask them to do. If you, for example, in your sheer genius, decide that a specific set of students get higher priority over another set of students on a specific set of classes, the program will do exactly that. It will not discriminate and will not make any distinction for you. For example, regardless of the fact that day and evening students alike, take OLA and elective subjects on equal footing, the program will not recognize that until you tell it to.

They should have properly tested the program before hoisting it upon an unsuspecting studentry. Next time, they should also publish the results. Meanwhile, I have to hope and pray that I don't have to go through this shit ever again.

No comments: