Electronic Transactions on Artificial Intelligence (ETAI)

Bibliographic Resource for Area Editors


The purpose of the bibliographic resource is to offer a set of tools whereby area editors can administrate, as easily as possible, the various kinds of bibliographic references that they need to deal with:

In a longer range perspective, we would like to offer these bibliographic services directly to the constituency, e.g. as a way of facilitating the administration of reference lists in research articles. See current official presentation of these intentions.

The basis for the present resource is formed by the contents of the LIDOS database which has been built up at the DFKI, together with software developed in Linköping. Technical software aspects will be described elsewhere; the present page describes the view on the bibliographic base that is being offered.


Indices of conference and journal literature

From the database, we are able to generate tables of contents for conference volumes and journal volumes in a uniform fashion. One straightforward way of extracting this information from the database for inclusion in HTML-based text (such as ETAI News Journals) is to cut-and-paste from the files offered here.

Please click here for a menue of conferences where this information has already been set up.

Disclaimer: There are still some minor bugs in these files - editor and publisher of conference volume missing, missing accents on author names, etc. Apologies to all concerned, but the present is only for illustrating the idea; bugs will be ironed out successively.

Here, the "indlis" index is intended as raw materials when an area editor wishes to select a subset of the articles at a particular conference, and present them to her/his readership as articles which are of particular interest to the area. The "indbib" index is similar except that each entry in the index contains the mention of the conference in question. Therefore, "indbib" is appropriate if one is preparing a mixed list, containing articles at several different conferences/journals, whereas "indlis" is appropriate if one puts several articles from the same conference under a joint heading.

The "indreg" index presents the authors at the conference in alphabetical order, and with clickable references to their respective articles. This index requires a web browser that supports frames. Note that alphabetical column in the rightmost window, which can be used for selecting a certain section of the author list.

The "indbib" index, finally, is the same as is presented in the author window of "indreg"; it is presented separately in order to facilitate making a printout of that list. Note that the articles are marked as clickable also under "indbib" (actually, the same HTML file is being used), but clicking without a frame context obtains no effect (I suppose). (An improved layout that adapts to the existence/nonexistence of frame support is foreseen).

In all cases, the articles are identified with an identifier such as c-ecai-92-165, using the same notational scheme as in LIDOS, and where the last part refers of course to the starting page of the article. Although the scheme is fairly obvious, its full syntactic definition of this scheme needs to be worked out. (There is a soft spot if the same conference is repeated several times during the same year, as has been the case for WWW).


Indices of recent conferences and workshops

The coding scheme shown in the previous section does not work at the point where the program for a conference or workshop has been advertised, but the proceedings are not yet available, since the page numbering is not known at that time. The following example, albeit for an older workshop, illustrates the use of author names as an alternative coding basis. (However, this method also has its disadvantages, since it runs into problems if several papers have the same first author).

EWSP-91: [ indlis], [ indbib].

In a case like this, with less than a dozen articles, it is of course not meaningful to make an author index.


Register of A.I. authors

The register of A.I. authors should list the following information:

(The page obtained by clicking contains a presentation that has been generated from the register; the register itself is a database. The same disclaimer as above applies also for this register).

It is intended to use this information in order to equip each mention of a registered author's name, in the indices described above, with a hot link to his or her home page. As a secondary effect, one should also be able to obtain more uniformly correct presentation of names than what is practically possible without this database support.

It is debatable whether, in the absence of a WWW page, one should equip the author name with an E-mail reference. My inclination would be to not do that, unless the author specifically requests it, but this may be discussed. Anyway, it is useful to have the E-mail addresses in the register in order to be able to reach the author at all.

The register page mentioned above is not intended for public consumption; it is intended as a way for the area editors to know which names are in the register and which are not. The present contents of the register consists mostly of researchers in the area of actions and change, but it would of course be very desirable to add researchers in other areas to the register.

I foresee that each area editor may wish to maintain a more detailed register of/for researchers in his or her own area. I would like to keep the general ETAI register as simple as possible, and restrict it to the kind of information mentioned above, but it would also be appropriate to equip it with references to that (those) ETAI areas where the author is also registered and known.


Caveats and future plans

The contents of these registers, and in particular the presentation of authors with names of unusual syntax is sometimes incorrect or imperfect. It is intended to improve this gradually, both by editing the database and improving the software.

The transformation from LIDOS entries to the present database is fully automatic, except that minor flaws are sometimes generated, so it would be technically possible to add several other conference series. (In fact, the IJCAI and AAAI listings were computed completely mechanically, without any manual intervention. It was done with a sequence of commands, but it might as well have been one single command per conference). However, I would like to have feedback about the proposed structure before it is extended to very many conferences.

Once we have an agreed structure for past articles in conferences and journals, it will be very important to accomodate *recent* conferences, journal issues, and workshops in a quick reactive mode. This will require both some additional software efforts and a certain cooperation in order that the data entry can run smoothly.

Naturally, it will also be natural to add direct links to the respective articles, in all cases where they are available on-line. A certain care has to be exerted, however, so that we do not violate any copyright rules. The same holds, in fact, for the so-called collection rights: Using "indlis" and "indbib" as sources for the manual construction of specialized bibliographies is OK, but it might not always be allowed to put these lists on-line in extenso, particularly not if the lists are equipped with links to the full texts of the articles.


Latest update: 20.2.1998
Administrated by Erik Sandewall, Linköping University, Sweden. E-mail ejs@ida.liu.se.