Ranking of CS Departments Based on the Number of Papers in Theoretical CS | Hacker News Reader

Ranking of CS Departments Based on the Number of Papers in Theoretical CS

projects.csail.mit.edu

24 pointsradmuzom14 comments

This is just plain stupid. Stop upvoting this kind of links.

There are enough nonsensical ranking and index computed on meaningless quantity, we don't need more of it. I'm not even talking about the pragmatical aspect of why this specific ranking is dumb (quantity over quality, comparing department of different size in number of people without taking it into account, many very important conferences in theoretical CS are missing (POPL, PODS, CCS, …), there are also journals missing, etc.).

> Our method for ranking is very simple but robust.

Except it uses the ranks of conferences, which is computed based on more similar bullshit.

Moreover, what is the point of such a ranking?

I really don't know. For some reason we are infatuated with ranks. Obviously they mean something -- but they don't mean everything. But I think to rely on them solely would be misguided.

A graduate student is looking for a good fit. For a graduate student picking a graduate school rank should be a factor but only in generating candidates for applications. Beyond that a wise student would investigate the department to see which is the best fit for him irrespective of the schools ranking. This is especially true if the graduate student knows the area of research he is going to commit to.

For a grad student looking for a PhD subject and advisor, taking such rankings into consideration is just mad. Doing a PhD means that you enroll for a multi year relationship with both your subject and your advisor. Choose qualitatively, with your mind, your heart and your gut, not quantitatively.

> Disclaimer: If you find the ranking in this website offensive, please ignore it. This data is not official at this point.

I, for one, found the data interesting. I didn't consider it to hold any meaning even before noticing the disclaimer, though.

I don't think it is offensive at all. Just stupid. And politically dangerous. It is hard enough to battle against this kind of rankings in academia, we don't need more.

Shouldn't they use the ratio publications/faculty size instead of absolute numbers?

Wouldn't that depend on how you want to apply the results? For some things, absolute quantity might be the important measure.

Over the past several months we have developed a ranking based on a hard, measurable method for the top 50 U.S. universities.

"Hard, measurable" != meaningful.

And counts of articles published are especially meaningless as measures of the quality of either the research actually done in these departments, of of the overall educational experience at those places. It's like measuring developer quality by LoC committed, or the number of JIRA tickets they've closed.

Quantity over quality?

Should be "based on the citations by members of OTHER faculties"...

Or... doesn't that happen nowadays ?

(everything else will just blow the system ... apparently)

Most institutions rate articles by impact factor for purposes of determining tenure/promotion, which is derived from the journal, which in turn is derived from the number of citations (the idea is that if a paper got into journal X it will probably end up being cited roughly as often as other papers published in journal X), so the sum of the impact factors of the articles would be a reasonable compromise (along with an average per faculty member).

Research about research. Always seems like the time would have been better spent on actual research.

US only? :(

Apparently. And, amazingly, a comparison by MIT professors that ranks MIT first.

Reply on news.ycombinator.com