DJIN: Detection in Journals of Identifiers and Names
Résumé
We dreamed of it, we developed it, and now we use it.
DJIN is a powerful tool that recognizes astronomical object names in full texts.
DJIN is very efficient and helpful for the SIMBAD team who have been dealing with an ever increasing number of astronomical articles.
DJIN detects most of the astronomical object names quoted in full-text articles, but the team still has to check and validate the names, to deal with new identifiers, to verify cross-identifications and to update SIMBAD with new astronomical data (position, magnitudes, etc.). That is, the team work is concentrated on value-added aspects, the best use of the team's expertise. This was an important consideration in the design of the software.
DJIN provides more than just the recognition of names; it says how many times an astronomical object is cited in a text (whatever its identifier is), where it is cited (title, abstracts, keyword, tables, figures, text, etc.) and keeps track of the relation between the identifiers and articles.
DJIN is fully integrated in the SIMBAD process, and interfaces the updating software used daily by the team. It is also a starting point for new features like linking SIMBAD and NED, and computing the relevance of each paper attached to one object.
DJIN has been fully tested by the whole team to check both the quality of detection and the tool's ergonomics. Team feedback has been critical for the success of this difficult and risky endeavor.
In this paper we describe this tool, and our experience after two years of usage; we discuss also the the significant changes in our daily work that DJIN has triggered.