3 The CDS activities

CDS activity has different aspects. Some are directly visible to the users, whereas others, though fundamental for maintaining the CDS expertise and role, may be less conspicuous.

The most perceptible activity of CDS is certainly the development, maintenance and on-line diffusion of reference, value-added databases and services, such as SIMBAD, VIZIER, ALADIN, the Dictionary of Nomenclature of astronomical objects outside the solar system (Lortet et al. [1994]), the AstroGLU discovery tools (Egret et al. [1998]), etc. The CDS services are described in more detail in Sect. 4, and in the set of companion papers by Wenger et al. ([2000a]), Ochsenbein et al. ([2000]), and Bonnarel et al. ([2000]).

From the point of view of contents, CDS deals with selected information: raw observational data are generally not available at CDS, but rather upper level data such as observation logs, catalogues, results, etc. This "reference'' information is then documented, organized, and made accessible in the CDS services.

In addition, in order to cope with the congestion of inter-continental connections, CDS has developed an active policy of mirror copy implementation. The redundant availability of data on several sites is also important to ensure data security. Mirror copies of SIMBAD and ADS are installed in CfA and CDS respectively (Eichhorn et al. [1996]). Mirror copies of VIZIER are installed at NASA ADC and NAOJ ADAC (which also hosts a copy of the Dictionary of Nomenclature), another one is foreseen at the Indian Data Center (IUCAA, Pune). CDS hosts mirror copies of several electronic journals and of CFHT documentation.

Less visible from the users, but an important field for networking international partnership, CDS develops generic tools and distributes them to other information providers: for example, the GLU (Générateur de Liens Uniformes), for maintaining links to distributed heterogeneous databases (Fernique et al. [1998], Sect. 5.1), or the SIMBAD client/server package, which allows archive services and the ADS to use SIMBAD as a name resolver.

CDS is also active in the development of exchange standards, such as the bibcode, first defined by NED and the CDS (Schmitz et al. [1995]), and now widely used by the ADS and the on-line journals, or the standard description of tables, defined by CDS and shared with the other data centers and the journals (Ochsenbein et al. [2000]). CDS collaboration with journals is described in Sect. 5.2.

CDS expertise in the domain of astronomical data is also useful for projects. Strasbourg Observatory has been deeply involved in construction of the input and result catalogues of Hipparcos & Tycho ([1997]), and SIMBAD has been used as a basis for ROSAT identification tools. At present, the XMM Survey Science Center relies on the CDS services to build a cross-identification database for the X-ray sources observed by the satellite. The CDS participation to the DENIS and TERAPIX/MEGAPRIME surveys is described in Sect. 5.4.

Finally, to keep up with the pace of technical evolution, CDS has to develop a resolute activity in the domain of technological and methodological watch, and to undertake Research and Technology (R&T) actions to assess new techniques. The GLU development (Fernique et al. [1998], Sect. 5.1.), and the ALADIN Java interface (Fernique & Bonnarel [2000]), are examples of R&T actions which came out as operational services. More recently, one can cite the ESO/CDS Data Mining Project (Ortiz et al. [1999]), or the assessment of commercial object oriented database systems (Wenger et al. [2000b]) (Sect. 5.4).

In practice, the main challenge in the CDS activity is to constantly tune the contents and the services to the rapid scientific and technical evolution, to be able to deal with ever increasing volumes of information, and to be ready to respond to the new projects and to the evolution of the policy of the national and international Agencies. The guidelines in prioritizing the tasks are several: offer the best service to the users, ensure the quality of contents, and make the best of technological innovation. This implies that new developments are begun neither too early, to rely on techniques which are as secure as possible, nor too late, to offer the best possible services and improve the functionalities - hence the importance of technical watch. A balance has constantly to be kept, between very long term activities, on time scales of several ten years, to build up the contents; the development and maintenance of database systems and user interfaces, on time scales of a few months to a few years; R&T activities, on similar times scales; and operational constraints on a day-to-day basis. Hence the importance of careful strategy definition and activity scheduling.