Re: new version of high level language

From: Clive Page <cgp-at-star.le.ac.uk>
Date: Fri, 28 Feb 2003 09:08:15 +0000 (GMT)


On Thu, 27 Feb 2003, Reagan Moore wrote:

> We have the technology to map from an XML specification of
> attributes, values, and operations on values, to the SQL required by
> a particular database (Oracle, DB2, Sybase, SQLServer, PostgresSQL,
> Informix).

> The technology does require the registration of the table structure,
> foreign keys, and schema of each database for the automation of the
> SQL generation.

> Are there other types of databases in current use?

That depends what you mean by "types". Quite a number of well-used astronomical data archives don't rely on commercial or even free RDBMS, they use an astronomer-written DBMS. The LEDAS facility here, for example (ledas-www.star.le.ac.uk) now supports searching of USNO-B1.0 - but this isn't loaded into any DBMS, but uses the WCSTOOLS software written at CfA which access the USNO-B data files in their distribution format, and unpacks them on the fly. Seems efficient enough to me for simple cone searches (we don't yet support cross-matching with other catalogues). Other places do something similar: when I asked Francois Ochsenbein if he had managed to ingest the whole of USNO-B1.0 (over a billion rows) into Sybase he told me that it was too large and he hadn't even attempted it. I don't know how Vizier serves it up (maybe Francois can tell us) but I expect CDS wrote the software in-house. So these systems don't speak SQL at all.

There are also one or two facilities using object-oriented DBMS - we use O2 here for our internal XMM-Newton operations, and are about to make its interface public. This speaks OQL - semi-standarized, but not the same as SQL.

By the way, you left MySQL of your list of RDBMS - perhaps the most widely used one in astronomy?

-- 
Clive Page,
Dept of Physics & Astronomy,
University of Leicester,    Tel +44 116 252 3551
Leicester, LE1 7RH,  U.K.   Fax +44 116 252 3311
Received on 2003-02-28Z10:14:15