Wednesday, June 06, 2007

arXiv User Survey

Just a quick link: in case you ever wanted to express your opinion about the arXiv, take their poll (from June 4-8)

http://library.cern.ch/poll.html

It will take approximately 20 minutes, and has plenty of comment options to complain about the new arXiv listing (or the eternal bug in the search field if you search for a tag containing the word 'not').

Other points that I find worth mentioning: the arxiv should allow comments on papers, and a ranking (different from times cited). Comments would be helpful to avoid the increasing amount of 'reply-to-reply-to-reply-to's, ranking I would find a good idea because it's become almost impossible to find a good review or lecture notes if one doesn't know the author (and lecture notes don't usually become top-cites).

9 comments:

  1. The absolutely positively very best arXiv gravitation paper, ever,

    http://arXiv.org/abs/physics/0205089

    Table at end of paper. Turn a dial and a 60.5 kg lump exhibits 41 kg of buoyancy. WOWSA!

    ReplyDelete
  2. As per your last comments, you might be interested in scirate.com, a little toy of a website I wrote for implementing rating and commenting for arxiv papers.

    ReplyDelete
  3. Dear Bee, citation counts are imperfect - like everything in the real world - but any style of amazon.com-like system of helpful votes etc. would produce complete garbage.

    There is already a huge experience with it. When outsiders get the opportunity to organize campaigns about things they don't know, they will do so. Most of these votes would have nothing to do with the actual content of the papers.

    If something like that had to exist, it would have to be a Google-style ranking system including PeopleRank analogous to PageRank that would reflect the power of the voters. Effectively, e.g. Witten should have 1000 times more influence than e.g. Woit as reflected by his own votes or citations - which would ideally give you the same ratio, as a matter of correct calibration of the new algorithm.

    Anything that would include votes selected by Witten as less than 1000 times Woit would be bound to lead to complete junk.

    ReplyDelete
  4. Dear Lubos:

    This would never work in practice. The easiest thing to do is let people comment with their name (say, you need to be registered at the arxiv/spires database) and let everybody decide for themselves if they find the opinion of person x worth reading. a simple ranking of papers would at least show up the papers of interest (whether in the good or the bad sense). I definitely don't suggest something like a quality ranking (nothing could be further off my mind, it would be a disaster) - But it would be handy to find the basic and introductionary papers/reviews in a field more quickly.
    Best,

    B.

    ReplyDelete
  5. Simpler solution about classification - let the author add from a set of tags - lecture notes, expository article, Ph.D. thesis, regular research paper, summer school proceedings, etc., and self-classify their papers.

    I'm not sure I want my public library catalog either to carry comments by people. It might be nice to have a common independent site where people can comment on papers, and hold discussions.

    ReplyDelete
  6. Hi Arun,

    I probably didn't make myself really clear, sorry. I don't want a general discussion or comments to a paper. I just want a possibility to submit these comment-to papers directly as a reply to a paper and not as a new paper in the listing. (whether or not that has to be as pdf/ps is another question). It's that I have the impression that it happens more and more often, and it would be better to arrange it in a tree than in an every-day listing.

    I actually think there is something like a forum or a blog where arxiv papers are discussed, I was there sometimes (sorry, can't find the link). It seemed to me like a nice try, but didn't really work. I think the problem is basically what Lubos also pointed out: to keep the standard at least at arxiv level, you'll need to have some restriction on the user registration. I mean, everybody can discuss every paper on a blog, or in an arbitrary forum if they like. The question is the expertise of the audience.

    Best,

    B.

    ReplyDelete
  7. Ah, so 0705.2630 and its children, all linked together.

    Doesn't the "Cited by" help do that?

    Anyways, since you're in Rome, hope you're doing as the Romans do. :)

    ReplyDelete
  8. The ranking system you want should be achieved by talking to your colleagues... private comments are much more useful than public almost all the time.

    But nothing substitutes for wading through the arxiv - alas !!

    ReplyDelete
  9. If you are interested in a blog-like commenting system but
    (a) more permanent
    (b) a bit more formal / scholarly
    (c) integrated abilities to do other stuff that is handy for scientists, like uploading papers, including equations in posts, etc

    ...then check our nano.org which we've built to do that stuff. It's for any research broadly in the area of nanoscience. Join in if you're interested, we'll happily register any researcher as a 'scholar' so you can contribute -- email: invite at nano dot org

    ReplyDelete

COMMENTS ON THIS BLOG ARE PERMANENTLY CLOSED. You can join the discussion on Patreon.

Note: Only a member of this blog may post a comment.