ProData

This section overviews the ProData Prolog/database interfaces to ODBC, Oracle, Informix, Sybase and Ingres. The interfaces are also known as ProDBI when linked with Quintus Prolog. All PC versions are ODBC based as are the SICStus versions. All other workstation versions are individually tailored to one of the databases: Oracle, Informix, Sybase and Ingres, these are mainly available for Sun machines, but other platforms are supported. If you require more information, contact the vendor for the specific Prolog as given below.

The ProData interfaces are the result of many years research in the Prolog-database field. They have been engineered to an extremely high standard and have been tested extensively.

Keylink was the world's first producer of Prolog/RDBMS interfaces.

Currently the interfaces are marketed by:

• For MacProlog, 386 Prolog and WinProlog Logic Programming Associates Ltd Studio 4 Royal Victoria Patriotic Building Trinity Road London SW18 3SX Tel: 0181 871 2016 Fax: 0181 874 0449 E-mail: cspenser@cix.co.uk
• For the Poplog system. Integral Solutions Ltd Berk House Basing View BASINGSTOKE Hants RG21 4RG Tel: 01 256 55899 Fax: 01 256 63467
• For the Quintus and SICStus Prolog system. SICS Box 1263 S - 164 28 Kista Sweden FAX 0046 8 751 7230 Email: sicstus-request@sics.se

The above companies should be contacted for price and availability information.

Keylink has developed interfaces to many other databases which include: EMPRESS, Rbase, xBase, Unify and Mimer although these are not offered as products. Bespoke work is undertaken.

Clients include IBM, AT&T, Boeing, Rolls Royce, Logica, Unilever, British Petroleum, Shell, The Turing Institute and British Gas among many others.

Main features of ProData

The ProData interfaces allow ODBC, Oracle, Informix, Sybase and Ingres tables to be accessed from Prolog as though they existed within Prolog's environment as ground unit clauses (facts). This facilitates the use of Prolog rules over the contents of the database, with no need to download any part of the database. All database accesses are done on the fly, allowing the developed Prolog rule-base to sit alongside other concurrent tasks. Backtracking, cut, call, not and all other standard Prolog mechanisms work identically over the table accesses and the internal database, thus achieving the highest level of transparency possible. This software architecture allows a completely different style of database programming that leaves fourth generation languages way behind. All the benefits inherent in Prolog development can now be transferred to the database field without any disadvantages.

Using Prolog brings benefits in:

• Rapid development time from an interactive environment;
• Expert system style rules;
• Sound logical paradigm;
• Declarative programming style.

Prolog prototypes written for database applications no longer need to remain as prototypes they can now access the main external ODBC/Oracle/Informix/Sybase/Ingres database by the simple addition of one directive for each table accessed.

In particular the interface offers the following functionality:

• Concurrent access and update, with program control;
• Full data access transparency;
• Indexes used automatically;
• Any database function can be carried out from within Prolog, such as creating tables, adding/deleting records, allocating access rights; in fact whatever may be achieved via normal SQL;
• Automatic data conversions;
• Full access to the data-dictionary;
• Performance as SQL;

The latest version incorporates several significant improvements. New features include:

• Completely transparent cursor manipulation, there is never any need to reset the database cursors and predicates for cursor manipulation have been deleted.
• The ability to link any SQL query to a Prolog predicate name has been added. This will eliminate any efficiency problems as the full efficiency of the database is available for any query.
• It is possible to see the SQL generated for a Prolog access to the database by the use of specially provided predicates thus providing a powerful debugging aid.
• User configurable handling of NULL values.
• BLOB handling