Tuesday, May 09, 2006

Forward-Looking Links

A computer science paper goes through many phases: manuscript, technical report, conference submission and proceedings and journal submission and published version. As a paper goes through these stages they usually improve, adding more background and intuition, better and more detailed proofs and so on. If someone wants to read your paper, you'd like them to look at the latest version. How do you make sure that they even know about the latest version?

You can't go into everyone's paper proceedings and add a yellow sticky to your paper saying to check out the new and improved journal version. But in this electronic age we can, in principle, add these notes.

First of all keep the papers on your webpage up to date. Many people just go to an author's page to download a paper and often they find some ancient version.

But after that then what? ECCC allows one to submit a revised version or add a comment which could point to a revised version. arXiv allows one to add journal information to an existing paper. Both of these require actions by authors that rarely happen. The digital libraries of proceedings publishers ACM and IEEE-CS don't have any mechanism to add pointers to later papers.

The field should have some standard mechanism for updating pointers to future papers. Until then we have to rely on the readers to find the latest papers on their own and perhaps hope that paper search tools like Citeseer, Google Scholar and Microsoft Academic Search will point to the latest and greatest version of a paper.

4 comments:

  1. "The field should have some standard mechanism for updating pointers to future papers."

    What? Like trackback?

    ReplyDelete
  2. I think this is a good suggestion.

    ReplyDelete
  3. On a related topic, there is a new bill, the Federal Research Public Access Act of 2006, in the US Senate, co-sponsored by Joe Lieberman (D) and John Cornyn (R), that would require papers on federally-sponsored research be available for free on-line access within six months of peer-reviewed journal publication.

    You can read the press release and the NY Times article about it.

    ReplyDelete
  4. IMHO, that could really help to improve the research process. For instance, a new researcher could know better the real "state of the art" of the field he or she is interested in.

    ReplyDelete