Edinburgh Regional Computing Centre

18/10/76

1) Attend ICL IDMS User Group Meeting; 2)Discuss data management facilities on 2900 and exchange experience.

Minutes of the User Group meeting will be circulated within the next month,when they become available: a brief summary only is given here. The meeting was attended by representatives from MOD, BBC, GPO, HSA, BNF, UKAEA, Pirelli, ICL and other interested parties. Reports on the current status of 2900 IDMS implementation were given by ERCC, MOD, HSA and BBC. Several other representatives had recently implemented 1900 IDMS as well, with little difficulty.

A discussion was held on a working paper from MOD about proposed enhancements to the inline software of the next version of ICL IDMS (B150). A paper was also circulated giving - unofficially - ICL's planned enhancements. MOD confirmed that ICL had committed itself to supplying them with at least two of the major enhancements 'by mid 1977'. The discussion was fruitful, but technical. Copies of the paper discussed are available. Topics discussed included a fast-load utility (to be implemented in the next Cullinane release), and the need for a report generator/queue language. With the exception only of myself and the ERCC representative the User Group favoured an interface between IDMS and the FILETAB system.

Reports were made on the European and U.S. IDMS User Groups. A representative from Cullinane Corporation reported on the develop­ments planned for their IDMS, and confirmed that compatibility with the ICL product remained a high priority.

Unfortunately few of the people I met at ERCC had had much experience of the Data Management facilities on 2900, nor did they anticipate that there would be great user demand for any but the more rudimentary of the DMUS, not, at least, until the package had been more rigorously tested. This view was shared by the others at the IDMS UG, in particular MOD Bureau West, one of whose representatives told me he had used the DMUS extensively but found the facilities too restrictive to be of much use. At ERCC, there was considerable enthusiasm for FAMULUS as a hierarchical file-handling package alternative to IDMS; the conversion to 2900 from the IBM version at UCL would be undertaken by the PLU.

I also managed to discuss in some detail a proposed application of IDMS at Oxford with Dr. G. Stacey, which was of great use to me, and hopefully to the user concerned as well.