Counting PC-Authored Papers

---

This script relies on DBLP to pull the papers related to a given conference and year. There are three things the user needs to set for each conference. Let me use USENIX ATC as an example.

1. The conference name as seen by DBLP. For example, for ATC this is usenix.

2. The venue as seen by DBLP. Each paper belonging to this conference will have this venue set. For example, for ATC this is USENIX Annual Technical Conference.

3. The year of the conference.

You might need to poke around DBLP a bit to get these values. For example, if you go to http://dblp.uni-trier.de/search/publ/api?q=/conf/eurosys/2017, you can see what the venue and conference name parameters are for Eurosys.

PC Information

Unfortunately, there is no automated way to get the program committee information. You need to manually get the list of PC members and put them in a text file. You then process the list of PC members and add them to the pc[year][conf] list in checkpc.py.

An effort must be made to make the names of the PC members identical to how they appear on DBLP. For example, if an author is called "A B. C" on DBLP, their PC entry must be changed to reflect that (often it will be just "A C" in the PC list).

For authors with common names, their DBLP name might have a number. For example, "A B 001". The PC entry will have to changed to reflect the correct author.

Checking PC Authorship

After parsing the PC list and populating pc[year][conf] you must call check_pc(conf, year, DBLP-conf-venue). conf and DBLP-conf-venue must match what DBLP sees for the conference as mentioned above. The conf used in pc and check_pc must match.

check_pc will then print out the total number of papers, the number of papers with at least one PC member as author, and the percentage of total papers authored by PC members.

Same Output

usenix 2017
Total Papers: 63
PC-Paper Count: 10
Percentage of PC-authored papers:  15.87

Caveats

This scripts works to the extent the names of the program committee match their DBLP name for their papers. As such, it is likely it does not work well for authors with unicode characters in their names, and authors with common names. Thus, the results of this script are a lower bound on the number of PC-authored papers in the conference.