IBM last week announced details of how 8100 users will be able to maintain at least some of their investment in applications via the 9370 – see p3 alongside – but if the UK users we talked to are any guide, it has come much too late. In the UK most 8100 users seem to have already decided that there is little point continuing to invest in the software for IBM’s first ever distributed offering, although some users do plan to migrate what they’ve got onto DPPX/370. George Goodwin, secretary of the IBM Computer Users Association said that of the group’s 1,250 corporate members only one remains listed as a user of the 8100. And UK computer users have shunned the 9370 too, opting instead for IBM’s mid-range Systems 36 and 38. Goodwin said British users have have adopted their usual approach to computing tinker with the software to maximise functionality rather that invest in ever more hardware as their American counterparts tend to do, partly because of the difference in labour costs on either side of the Atlantic. Tesco Plc, once a loyal and avowed pioneer of the 8100 machines, does echo the worries of the German community (CI No 886), but instead of waiting it out and seeing what IBM can manage on the 9370, it has switched to a tried and tested operating system to match the architecture, DOS/VSE. Vote with their feet It could even be that users all vote with their feet away from DPPX, and DOS/VSE gets a new lease on life as a distributed operating system – which would be ironic since IBM has been trying to de-emphasise VSE on the 9370 in favour of VM. But it did envisage something similar when it offered a cut-down, packaged version of VSE for the 4321 in 1981, calling it SSX: perhaps the idea was a little ahead of its time. Tesco says it decided to move to VSE because it was not prepared to endure the continuing uncertainty about IBM’s plans for DPPX on the 9370 and admits that it would rather have had a growth path that did not involve the need to replace its operating system, thereby threatening its existing investment in applications – which sounds as if an earlier decision by IBM would have saved the day for DPPX at Tesco. The supermarketer intends initially to replace eight 8100s used for stock control at its depots. It has a further 10 to 20 8100s which will also have to be replaced, but Tesco has long had plans to install a company-wide point-of-sale system that was once to have had over 500 8100s at the heart of local processing for store managers. Now any developments along those lines will either centre on the 9370 or more likely on the machines in that UKP50m order for computers and terminals with which Nixdorf Computer blacked IBM’s eye earlier this year (CI No 890). Another 8100 user, Grants of St James, admitted that it was locked into the IBM distributed processing environment because it cannot afford to abandon the investment made in developing applications under DPPX. Steve Norbury, technical services manager at Grants said: We have developed applications to run under DPPX and it would cost us a lot of money to rewrite them. Grants has eight 8100s which it uses mainly for order processing. Very concerned Norbury conceded that the signals from IBM concerning the future of DPPX were far from clear. But he is confident that typical Cobol applications under DTMS on the 8100 will run on 9370 DPPX with minor modifications: If that’s not the case we will be very concerned, he said, conceding that attempts to ferret out more information from IBM had been unuccessful. Norbury is particularly keen to find out if IBM will decide to push its Systems Applications Architecture interfaces at the expense of Cobol and PL-1 on the 9370. We are happy with our operating system but we are not keen on the 8100. Perhaps deciding to go for distributed processing was not a good idea. The arguments we used to make it are not as valid as they were then, Norbury said. In fact he suggested that given subsequent developments, the classic arguments for distributed processing – poor quality telephone links between terminals and main
frames, resilience in face of hardware failure and the danger that the data processing department might hold the company to ransom during a strike have lost their force. Grants has spent heavily in DPPX, with UKP750,000 on hardware alone but That money is not the issue. We have invested 20 man-years in developing software. Other users are also convinced that IBM does not currently have appropriate technology for distributed networking, but underlying this, the idea of distributed computing itself is challenged. Perth, Scotland-based insurance company General Accident is planning to abandon its 8100 within the next two years. We think we should go totally centralised because it’s a wee bit painful if you have all your data spread out here, there and in effect everywhere. It makes data management and collecting management information that wee bit hard, says technical support manager Alastair Macpherson. When the company acquired its 8100s eight years ago, Macpherson said that decision was made because it could not find a large enough mainframe for all its applications. But why wait two years before throwing out the old machines? We want to get as much as we can out of them. DPPX is a joke Macpherson was also unimpressed by the 9370: DPPX is a joke. But I wouldn’t enjoy MVS on the 9370 either. It’s too small a box. 9370 is not even XA compatible so it won’t run MVS/XA and nobody uses MVS/370 at all these days. MVS/ESA isn’t that far away now and it hasn’t got a hope in hell of going onto the 9370, he said. British Steel was traditionally a big 8100 user, but has been getting rid of some over the years. The company, reluctant to talk about its computing strategy because of its relationship with the Department of Trade & Industry, said it has nine 8100s in operation under DPCX after getting rid of quite a few. Why? An insider said the company intends to jettison all its 8100 by year-end because they offer poor price/performance, but would not say what would replace them. Confectionary maker Rowntree Macintosh was happier to talk about its run-in with DPPX. Failure of its attempts to establish the extent of IBM’s commitment to DPPX convinced it to stop purchasing and development of applications on 8100s. It had six machines before deciding that DPPX was ‘going nowhere’ and plumping for DEC VAX machines – and won’t reappraise its decision in light of the 9370: The 9370 assumes there is a father node. That means we can’t guarantee that each of sites our sites can work autonomously if it needs to, it said. So where does this leave IBM? 8100 sales were disappointing throughout its life, and DPPX investment is certainly not the lifeline that will keep the 9370 from the same fate. Which means most 9370s must go into companies with an IBM mainframe but no experience of distributed processing with IBM hardware. And that’s the VAXkiller?