On 0xDE I learned that Google has a new serivce, called Google scholar citations. You can make a profile and collect all your publications that Google scholar can find. It also calculates several of these statistics about your citations like the h-index.

Using this one realizes a few things quickly:

  • Google scholar is not even close to be a reliable data base for mathematical publications. I find talks, poster, preprints multiple times,…
  • Google scholar is not even close to have a proper database of the “citation network”. Some preprint series (in which I have some preprints) include a list of all paper in that series at the end of each paper. Google scholar takes this as the list of reference and hence, produces wrong results here.
  • Its hard to tell different people with the same name apart.

It seems that one purpose of Google scholar citation is to improve on all three parts by letting the people doing the work (what is not a bad idea). Note that similar services exist for several disciplines. For mathematics there is MathSciNet which is maintained by the Amarican Mathematical Society. They really put effort in their database with respect to all three point above. Moreover, there is the Zentralblatt.

While it is really nice to have a search engine for science which also finds preprints and not only published material, this service makes it even easier to produce stupid things like “impact statistics”, “productivity indices”. On the positive side it is really helpful to see what paper do cite another one (this makes it easier to find the source of some idea or original proofs). On the downside, this service may lead to the impression that research output (if there is such thing) can be measured by numbers in any way. I do not think that this is true. In the end: What does it say that one paper has been cited 50 times and another one five times? Think about it. It does not mean that the first paper has been read more, had spread more interest, had more impact or is even better in any way. It plainly says that the first paper has been cited ten times more. That’s it. There are papers which introduced great ideas in an awkward way and then a follow-up explained the same idea better. Hence, the second one gets cited although the first is equally important… There are weird paper which get cited a lot because everyone does. There are paper which many people did not read but cite because the have heard or read that this paper is the reference for something.

By the way: My profile is here.