How remote access works

This section outlines how the remote access mechanism works. It is not strictly necessary to follow it in order to use catremote, though it may help you to appreciate some of the reasons behind some of catremote's behaviour. The configuration file used by catremote is no more than its name implies. It simply defines a list of remote catalogues and provides some details for each: its computer network address, the sorts of query that it will accept, a short description etc.

For every remote catalogue listed in the configuration file there must be a server running on a remote machine. This server will accept queries sent from catremote, interrogate the relevant catalogue to select the objects which satisfy the query and return the selected objects to catremote.

There is a standard protocol for both the queries and the returned results which allows catremote and the various servers to communicate. This protocol is a subset of a proposed general format for exchanging information between remote astronomical information services which is being developed at the Centre de Données astronomiques de Strasbourg (CDS) and elsewhere. The proposal is described in the working document Astronomical Server URL by M. Albrecht et al.[1]. It is important to realise that this protocol is general, and allows not just catremote, but also various other clients, such as GAIA[12] and SkyCat, to communicate with various different servers for differing purposes. Thus, it is not optimised for catremote, resulting in some peculiarities in the catalogues of selected objects written by catremote (see Section , below).

The Astronomical Server URL protocol, as its name implies, uses the Hyper-Text Transfer Protocol (HTTP) developed for the World Wide Web. Thus, in order for catremote to work successfully your local computer must be configured for running Web clients (such as netscape). Of course, most Starlink nodes (and, indeed, most networked computers) will be so configured. One way of thinking of catremote is that it is functioning as a very specialised Web browser. Similarly, the remote servers are, strictly speaking, `gateways' using the Common Gateway Interface (CGI).

There are various types of remote servers: catalogues, name servers, data archives and image servers. All are `catalogues' in the sense of returning a table of values. A catalogue server returns traditional columns such as position, magnitude, colours, spectral type etc. A name server primarily returns columns containing celestial coordinates and alternative names for the object. A data archive will return a normal catalogue of values but at least one column will list URLs pointing to images or `bulk data' files for the objects tabulated. It is important to realise that though catremote can return these special columns CURSA contains no facilities for interpreting them. When catremote displays the list of accessible catalogues it includes the type of each (catalogue, data archive, name server or image server) immediately after the name and before the description.