For a school with one of the best CS departments in the world, Stanford has a hilariously bad track-record with homegrown software. The new Axess is far from perfect, and its predecessor was nothing short of laugh­able. (Who ever heard of a website that was down 4 hours a day?) Coursework seems to do its job, but it costs somewhere between 10 and infinity times more to maintain than comparable systems elsewhere. Howev­er, these examples pale in comparison before Stanford’s masterpiece of malware. I speak of the single most evil computer program ever conceived: the Draw.

For the uninitiated: in order to get housing for the upcoming year, Stanford undergrads start jumping through the hoops each April for a process known affectionately as the Draw. First, one leverages one’s ethnicity, language skills, interests, and fake-interests to obtain a set of house-priorities. Second, one forms a group and is assigned a number. Third, based on how good one’s number is (and a careful examination of historical housing placements, and the priorities one has obtained, etc.) one submits a ranked list of house preferences. Then, all the data is plugged into a giant, secret computer that churns through it and a week later tells you where you get to live.

The good news about the draw is that Stanford guarantees housing to everyone who wants it, so whatever happens, you still end up with a decent room in a decent building in a beautiful place. The bad news is that unless you’re an athlete, disabled, Greek, or married (none of which provide you com­plete protection) the program will determine your housing assignment mostly by voodoo, which is acci­dent-prone and full of inherent unfairness.

What’s to be done? Good proposals fall into a few categories: improve housing, improve the rules, improve the software, tweak the software’s param­eters, and quit whining. Let’s consider each in detail (except the latter, which we’ll flout).

OK, it’s not really a fix to the draw itself, but Vice Provost Bravman recently proposed a major overhaul of undergraduate housing, the gist of which was that Stanford wanted to spend a lot of money to improve the material quality of life of undergraduates. The plan, despite being largely unformed, has so far been tragically maligned.

A campaign of detractors lead by (you guessed it) ethnic theme dorms has been passionately expressing vague fears about any deviation from the status quo. Not to state the obvious, but if the quality of hous­ing improves, we all win regardless of how the Draw comes out. That should be a lot more compelling than worries that future Ujamaa residents might not like FloMo dining.

Speaking of theme dorms, they’re what I have in mind when I say “improve the rules”. The idea of creating themed communities is maybe not a bad one (though the argument has certainly been made that ethnically-based ones are naturally divisive), but it hardly ever works. Ask a few people who live as priority residents how important their themes are to them, and you’ll see what I mean.

Basically, theme dorms select a set of arbitrary attri­butes as worthy of special housing consideration, and those who happen to have them (or the willingness to fake them) benefit. Seriously, some people spend four years in Stern/Wilbur doubles, while native Italians can get prime Row real estate in unpreferred years. Humbio majors have a dorm, but not Econ majors or Music majors. French speakers have a row house, but Spanish speakers only have a dorm, and Hindi speakers have nothing. The solution here is not to have more theme houses to hit every special interest, but to stop discriminating in the first place.

Of course, the software itself is also to blame for suboptimal housing assignments. Without a lot of clairvoyance, it’s hard to rank houses in such a way that you get placed in your favorite open location when your number is picked. Even if you succeed, the in-house draw can screw you in any placement. An ideal system would work as follows: people in the same group will be assigned consecutive numbers, then when each person’s number rolls around, they’re directed to a website that shows every unclaimed bed left on campus, and they can pick their favorites. This might not be impossible to implement (although it doesn’t really seem likely to be done in that way), but we could certainly come closer than we do now.

Any number of small changes could also prevent a lot of unnecessary pain. Even just giving people the option of ranking 20 dorms – rather than only eight – would be an improvement. Or maybe we could only have to enter preferences once per group, instead of individuals having to get them right, on pain of getting accidentally split off. Or maybe we could split the two preferred years into a 1-1000 year and a 1001-2000 year so everyone gets a chance in the more exclusive houses. These specific ideas aside, the interface is definitely due for all sorts of updates.

The Draw may be a Stanford tradition, but neither an old one nor a great one. Let’s fix it before some­thing about lying on a priority application finds its way into that “101 things to do before you graduate” list.