We are Under Attack
We are Under Attack By the Least Publishable Unit
3 Evidence Zillions of mediocre papers — On increasingly narrow topics — Of interest to almost nobody — With high serial overlap Number of submissions to conferences (including this one) is out of hand — Requiring a PC of multitudes — Generating pretty random reviewing — With a goal of rejecting a paper if it is the least bit controversial
4 Evidence General wisdom Until acceptance { float a bad paper to a conference; improve it based on reviews; } Leading to yet more submissions
5 Further Evidence I got hired at Berkeley in 1971 — With 0 publications I got tenure at Berkeley in 1976 — With 7 publications You can’t get an asst professor job unless you have at least this Tenure cases typically have 4X this number Asst. professors are neither 4X smarter nor work 4X harder 35 years later
6 Two Modest Proposal Increase demand Decrease supply
7 Increase Supply Co-opt a few more conferences Anything with the words “big data” in the title is a candidate, e.g. — ACM Symposium on Cloud Computing — New IEEE Conference on Big Data Move major conferences to a poster format — At least for thenarrowly focused papers — So the 3 people who are interested can yack with the author
8 Decrease Supply Create a “Standing Committee” of “gray beards” — Say 20 people — Organized by subareas Which would review up to 2 extended abstracts per year — say 6 pages — from legitimate researchers (who would have to be screened) — I.e. take your best pair of shots per year Basically a universal program committee
9 Decrease Supply Every paper would get a score — Say an integer between 0 and 20 Any conference (PODS, CIDR, SIGMOD, VLDB, ICDE) could then pick the top ranked and previously unpublished abstracts (presumably with a sub-area bias) — Which could then be expanded into a full paper Has the nice feature that rejected papers automatically roll over to the next available conference — Could augment this with a mentoring system…
10 Over Time Gray beard average score (GBAS) would be the significant metric in tenure decisions…. — Importance not volume!