28
$\begingroup$

Other than ACM, IEEE computer Society, Google Scholar which is the best site to get bibtex entries for computer science related articles ?

$\endgroup$
  • 1
    $\begingroup$ related: 1 $\endgroup$ – user5109 May 17 '11 at 9:13
  • 5
    $\begingroup$ It would be great if someone were to make a community-edited site with accurate bibTeX entries for TCS. It seems like it wouldn't be too hard to set up. $\endgroup$ – Robin Kothari May 17 '11 at 22:48
  • 6
    $\begingroup$ @Robin: (1) The volume is huge: DBLP lists 1600000 papers. (2) Everyone has their own idea of what is the correct Bibtex entry. It is easy to point out that some entries are clearly incorrect, but it is not easy to agree on a single correct format. Journals are straightforward, but conference papers are tricky. (3) Even if such a site existed, would you be able to trust that the entries are correct, without double-checking them? If you need to double-check everything anyway, why not simply maintain your own database (with just those papers that you need to cite)? $\endgroup$ – Jukka Suomela May 17 '11 at 23:24
  • 2
    $\begingroup$ @Robin, there are already: liinwww.ira.uka.de/bibliography/Theory/index.html Some researchers also put their bib files on their website. $\endgroup$ – Kaveh May 18 '11 at 3:20
  • 3
    $\begingroup$ @Jukka, I don't think you need to double check everything, and you can check them as you use them in papers just the way you check your papers for errors you will also check the references part. I don't think that is a big deal. Finding and entering the data into your bib file is the harder part IMO. And I think most people will not care about the format of the entries too much as long as they are consistent. $\endgroup$ – Kaveh May 18 '11 at 3:25
34
$\begingroup$

You don't get correct TCS Bibtex entries from anywhere.

  • CiteSeer, Google Scholar, etc.: the Bibtex entries are garbage, worse than useless.

    • Examples: Many conference papers in Google Scholar are exported as an @article, with (some version of) the title of the book in the journal field. Google Scholar abbreviates the first names of the authors. And then of course we have ridiculous things like author = {Submission, H.C.F.} – Google Scholar populates the fields by picking some words from the cover page of the paper.
  • Publishers: the entries are a bit better, but you cannot rely on them – you must check every single field manually anyway. IEEE tends to be worst, ACM and Springer are little bit better, but even with the latter, you need to do manual editing and cross-checking. Springer has a strange idea of what is the title of a proceedings book. ACM gives book titles in a strange mixture of upper-case and lower-case letters. And, as usual, if there are accents in the authors' names, or any math in the title, all bets are off.

    • Examples of booktitle fields for conference papers: Springer might produce something like booktitle = {Distributed Computing} for a proceedings volume – it requires a lot of imagination to figure out that it actually means "Proc. 23rd International Symposium on Distributed Computing (DISC 2009)". IEEE exports unreadable titles such as booktitle = {Sensor, Mesh and Ad Hoc Communications and Networks, 2007. SECON '07. 4th Annual IEEE Communications Society Conference on}. ACM is usually fairly good, but you need to fix the mixture of upper and lower case letters: booktitle = {Proceedings of the twenty-first annual symposium on Parallelism in algorithms and architectures}.

    • Examples of titles with math: ACM might produce (\&\#948;+1) instead of {$(\Delta+1)$}. IEEE might produce Otilde(radic(log n)) instead of {$\tilde{O}(\sqrt{\log n})$}. I am not kidding you.

  • MathSciNet: high-quality Bibtex entries for journal articles, but the coverage of TCS is poor, and conference papers are not necessarily that well indexed.

    • "Core TCS" conferences such as FOCS, STOC, and SODA seem to be covered fairly well, but anything else is more patchy. For example, there seem to be few papers indexed from PODC or SPAA.
    • The entries of the conference papers are not perfect. You can find something like @incollection instead of @inproceedings, or proceeding books such as BOOKTITLE = {Distributed computing}.
  • DBLP: reasonably good, but once again, a lot of data comes from the publishers, and you need to double-check it anyway (beware of accents).

    • Examples of accents: Michal Hanckowiak instead of Micha{\l} Ha{\'n}{\'c}kowiak.

As JɛffE pointed out in the comments, the correct title of a conference volume is a matter of taste (and a matter of interpretation). For example, LNCS volumes may have useless main titles and ridiculously long subtitles; therefore even if you had pedantically correct bibliographic entries, you most likely would like to edit some of them slightly, for readability and for consistency.

But as soon as you start to tweak the titles of the conference volumes, it becomes obvious that even for your own purposes, there are many possible right answers. When you are running out of space, you might prefer "Proc. STOC 2010" to "Proceedings of the 42th ACM Symposium on Theory of Computing (STOC, Cambridge, MA, USA, June 2010)". This answer at the TeX site gives one example of how to deal with multiple versions of the titles, so that you can easily switch between different variants.

$\endgroup$
  • 1
    $\begingroup$ My favorite publisher example is Springer's “journal = "Discrete {&} Computational Geometry" ” $\endgroup$ – Jeffε May 17 '11 at 19:59
  • 5
    $\begingroup$ My least favorite aspect of ACM's bibtex is that they make the doi entry be something other than the doi (instead they use a url derived from the doi). Using series= for the conference acronym is also annoying. I wouldn't call MathSciNet's coverage of TCS "extremely poor": it's patchy, but not bad. Their coverage of CS that isn't TCS is extremely poor, unsurprisingly because it's out of scope for them. But they're still where I go first nowadays to look for good bibtex. $\endgroup$ – David Eppstein May 17 '11 at 21:50
  • 2
    $\begingroup$ @Jukka: Springer's BibTeX entries reflect the title that's actually printed on the front cover. Most LNCS proceedings volumes are not actually entitled "Proceedings of the #th International Symposium on (whatever)". Following strict scholarship, the BibTeX is correct; it's the front cover of the book that's screwed up! (Like everyone else, I abandon strict scholarship and (try to) use functionally correct titles for proceedings.) $\endgroup$ – Jeffε May 18 '11 at 3:41
  • 1
    $\begingroup$ @JɛffE: I know, and I hate it. Actually LNCS proceeding volumes have a subtitle that contains the full information. For example, the title of the volume might be "Distributed Computing", with a subtitle "23rd International Symposium, DISC 2009. Elche, Spain, September 23–25, 2009. Proceedings". By slightly re-arranging the information, you get "Proceedings of the 23rd International Symposium on Distributed Computing, DISC 2009", which is already fairly close to a functional title (and one can still guess that it is the same book...). $\endgroup$ – Jukka Suomela May 18 '11 at 9:49
  • 4
    $\begingroup$ Incidentally, I have seen LNCS books with full titles such as "On the Move to Meaningful Internet Systems 2006: OTM 2006 Workshops. OTM Confederated International Workshops and Posters. AWeSOMe, CAMS, COMINF, IS, KSinBIT, MIOS-CIAO, MONET, OnToContent, ORM, PerSys, OTM Academy Doctoral Consortium, RDDS, SWWS, and SeBGIS 2006. Montpellier, France, October 29 – November 3, 2006. Proceedings, Part II". I was going to give another example but these comment fields are too short. $\endgroup$ – Jukka Suomela May 18 '11 at 9:55
5
$\begingroup$

If you are looking for BibTex entries for articles/papers whose titles and/or authors are already known to you (i.e., you just need the BibTex entry), DBLP is my favorite choice and very convenient

http://www.informatik.uni-trier.de/~ley/db/

See, e.g., http://www.informatik.uni-trier.de/~ley/db/journals/jacm/jacm58.html

$\endgroup$
4
$\begingroup$

I don't know if these are the best option, but you can give them a try:

CiteSeer has some BibTeX entries. Moreover, specific publishers (such as Springer or ScienceDirect) offer an "Export Citation" option, in which you may choose different formats (BibTeX, EndNote, etc.) to export the citation.

Here's an example: http://www.springerlink.com/content/pleucejg0nlfna9m/export-citation.

One more note: Google Scholar does not let you to export the citation by default. You have to set the option in the Google Scholar preferences.

$\endgroup$
4
$\begingroup$

There are collected bibliographies like:

  1. ECCC Bibliography Data Bases page,
  2. The Collection of Computer Science Bibliographies.
$\endgroup$
  • $\begingroup$ Though sometimes cumbersome, I also like to use The Collection of Computer Science Bibliographies. The main reason is that it provides several bib entries on a single page. This usually makes it possible to construct a complete entry from several incomplete entries. $\endgroup$ – Don Sheehy May 17 '11 at 13:58
3
$\begingroup$

Collaborative bibliography managers such as Mendeley and CiteULike have been around for several years, but have not yet caught on in the theory community (at least, they include a vanishingly low proportion of the papers I would like to cite).

Note: I am listing these services in a spirit of completeness. They are relevant to the question, but I do not endorse any collaborative bibliography manager.

$\endgroup$
2
$\begingroup$

MathSciNet and the SCI are fairly standard resources (but are behind paywalls).

$\endgroup$
2
$\begingroup$

I attended a talk by Sebastian Lindner in early April this year. He works for Springer Materials working on normalising citation data (can't find a reference, sadly). This is still work in progress, but we will hopefully see some significant improvement.

From what I remember of the talk, authors can help a lot by adhering to some standards, wherever they take them from.

$\endgroup$
1
$\begingroup$

Bibsonomy is place where I try to look for bibliography information, when other services fails. Additionally it's good to collaborate by creating "friends" lists.

$\endgroup$
0
$\begingroup$

Odysci

From Odysci - How it works :

Odysci is a web portal for search and rank of technical papers and peer-to-peer collaboration. Using the Odysci portal researchers and developers can search for technical papers published by the main publishers in the computer science and electronic engineering areas, obtain ranked lists of papers and their links, and collaborate with peers by reading and posting comments about papers. In the future, Odysci will also display technical white papers by companies interested in divulging their technology.

$\endgroup$
  • $\begingroup$ I see no ability to obtain bibtex-formatted data from odysci. Maybe it is there but well-hidden? $\endgroup$ – David Eppstein Jan 4 '12 at 20:51
  • 2
    $\begingroup$ It's there but clunky. If you click on the little 'bib' link, you can "add it to a list". You can then view the list in bibtex format. $\endgroup$ – Suresh Venkat Jan 5 '12 at 6:23
  • $\begingroup$ btw. it's somehow similar to Bibsonomy's approach. When you are logged in, you put there into basket and then you can export whole basket into bibtex. Here is similar, but it's called "lists'. $\endgroup$ – Grzegorz Wierzowiecki Jan 6 '12 at 10:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.