volution des publications scientifiques Bibliometry and mathematics tienne GHYS

1
Évolution des publications scientifiques
Bibliometry and mathematicsÉtienne GHYS

• Académie des Sciences
• 14-15 mai 2007

2
Mathematics a small world

• ? 2 500 000 papers written since Euclid
• ? 70 000 mathematical papers each year.

3
(No Transcript)
4
Mathematicians dont write too much
but they read a lot !

• Typically, 1 or 2 published papers per year
  (averaged over the career) is considered as a
  good activity.
• A mathematical paper has usually 1 or 2 authors,
  sometimes 3.
• Usually, the bibliography of a paper contains 10
  to 20 items.

5
Number of Citations a useless tool in
mathematics

• Highly depends on the subdomain
• Example 27 mathematicians from the Académie des
  Sciences !
• From 5850 to
• 60

6
Mathematical  time scale  is long !

• Typically a paper is published 1 to 2 years after
  submission (sometimes more)
• Referee procedure is fundamental (and long!)
•  Half life  of a maths paper is 10 to 20 years.
  Longer for some very good papers.

7
The  Impact factor  is not appropriate for
mathematicians
8
Example the Comptes Rendus Mathématique

• Among the mathematical papers published between
  2000 and 2004
• 3 469 contain a citation to a Compte Rendu Note
  published in the 90s
• 1 600 in the 80s
• 871 in the 70s
• 465 in the 60s
• 190 in the 50s
• 123 in the 19th century !

9
How to evaluate a mathematician ?

• Only good method read his/her papers !
• Acceptable method trust somebody who read
  his/her papers.
• Prizes, Awards, Invitation to Congresses, etc.
• Comité National des Universités, Comité National
  dévaluation de la recherche scientifique, etc.

10
The mathematical world is small enough to allow
a  personalized  evaluation.
Read Condorcet ! Essai sur lapplication de
lanalyse à la probabilité des décisions rendues
à la pluralité des voix (1785)