RE: another scary use of csindex.com

Stefan Savage (savage@cs.washington.edu)
Wed, 30 Jun 1999 18:55:34 -0700

We gotta be a little careful looking at the tail of that curve because
presumably citations are a function of time and number of papers as well as
area popularity and paper quality. Papers from 91 have had 8 years to be
cited, vs papers from 97 which haven't have much chance (otherwise Eraser'd
be way up there ;-). However, the blip in 93 does look significant, so
perhaps you're right (maybe 93 was just a really bad SOSP ;-).

FYI, By my count the number of papers appearing at SOSP as been increasing,
not decreasing:

87 - 19
89 - 20
91 - 17
93 - 21
95 - 22
97 - 23

- Stefan

P.S. This is also only the citations found on papers accessible via the
Web. its not so surprising that there aren't too many papers cited before
year 6 B.W. (before web)

-----Original Message-----
From: Tom Anderson [mailto:tom@boris.cs.washington.edu]
Sent: Wednesday, June 30, 1999 5:54 PM
To: bershad@cs.washington.edu; lazowska@cs.washington.edu;
levy@cs.washington.edu; syn@cs.washington.edu
Subject: another scary use of csindex.com

I needed to argue for SOSP to accept a measurement/experience paper
based on its interest-level, so I did the following:

search "Symposium and Operating and (Systems or System) and Principles or
SOSP"

in other words, find the most referenced papers at a conference (ignoring
award papers, since those usually get referenced via TOCS).

scary results. on the plus side, we're 3 of the top 6 ;-)
on the downside/upside?, references to SOSP papers had been
growing exponentially, but aren't anymore ;-)

tom