Sunday, December 10, 2006

Ranking CS Conferences.

We are constantly trying to rank things. Mostly, we understand that there is a partial order, a tournament, or something more, but the world often seeks a simple total order. I saw a ranking of CS conferences: citeseer (2003?) based on normalized citations; cs conf ranking (every quarter, warning lights, uses quality of referee reports (?) and funding support for travel to confs for students (??)); the curated list at alberta (no Theory confs! ACM/IEEE confs make Tier 1, European confs make Tier 2); another curated list, this with Theory confs;...

Amuse: Saw a reference to *Googol* Page Rank system in the best practices memo at the CRA site.


Anonymous Anonymous said...

there are theory conferences in the alberta list (below the "regular" lists) - STOC, FOCS, ICALP, SODA, ISMB made the 1st tier.

5:39 PM  
Anonymous Anonymous said...

Rankings can be useful when consumed in moderation.

Say, when evaluating a candidate, it is safe to look at the overal strength of the venues in which s/he published her top 10 papers: eight in tier one conferences plus two in tier two is good, while eight in tier two and two in tier one conferences is not as good.

Using the rankings to compare a specific pair of papers, on the other hand, is just too noisy as to be of any use.

1:16 PM  

Post a Comment

<< Home