IBM announced Systems Applications Architecture in the UK just over a year ago, but for a company renowned for its marketing acumen something seems to have gone wrong because users are far from sold on SAA. UK users claim they see little evidence of IBM’s commitment to SAA despite the announcement that was supposed to herald the advent of easier and consistent computing across the IBM product range from the PS/2, through Silverlake to 370. Nothing has happened. All it is at the moment is all words and no action, Martin Skelton technical support manager at Zurich Insurance in Portsmouth said. Another user, Alastair Macpherson, technical support manager at General Accident in Perth Scotland said: It’s all extremely vague. SAA looks interesting but I am not sure exactly what it is. And Tom Drake of the IBM PC User group offered an equally disparaging view of SAA seen from the micro end of the computing community: In 10 years’ time someone might have a perspective on SAA – it won’t be around until then. But Muriel Jones, SAA programs manager for IBM UK quite naturally takes a very different different view. Contrary to signs of antipathy among users, Ms Jones said during the past year users have taken to SAA and are now busy laying the groundwork for the implementation of applications based upon it. SAA is one of the most significant announcements IBM has ever made. The next 20 years It’s really showing where we are going over the next 20 years. Users and software vendors are preparing for SAA. They are very keen to get all the information they can because they see it as where they are going and feel the sooner they start the better, she said. But IBM either refuses or cannot give a list of these enthusiastic users. The company expects to make two major SAA announcements each year and has already certified a range of languages and software enabling tools under SAA spanning the OS/2, VM/CMS and MVS/XA environments, including Cobol, Fortran, DB2 and communications products alone. A 350-page manual defining user interfaces for the common user access strategy under SAA, which aims to standardise man-machine interaction using the PS/2 based Presentation Manager windowing environment has been published and despatched to software developers and users, Ms Jones said. But why do users remain unconvinced that SAA has arrived? Asked to comment on SAA so far, Ian Stinson of Tesco said simply: When is it going to happen and how? A year on, users who care are still struggling to grasp what an SAA application is. Isn’t there more to SAA than the common user access, easier and less costly porting of software by having a core of specifications on which all major IBM software is developed using a common programming interface – languages, commands and calls? Yes. So what is an SAA application? Nothing is written in concrete as to what an SAA application is, IBM’s Ms Jones answered, perhaps illuminating why further down the chain users may feel that where SAA is concerned nothing has yet been pencilled in, yet alone chiselled. According to Ms Jones, guidelines as to what is an SAA application do exist: it should execute on SAA systems – PS/2, Silverlake, 370; adhere to the common user interface based on Presentation Manager; it must use a relational database with an SQL interface – preferably DB2 – and it must be integrated and expandable enabling additional hardware-specific functions to be added. What users should be doing now is to get ready for SAA by careful planning. They must decide on which features of the common user interface they will use; choose their programming language; look at relational databases and hire the experts to run them and through applications design decide which products they are going to use. Ms Jones feels this is just what is happening among IBM users. SAA watcher Michael Killen of Palo Alto, California-based Killen & Associates reckons that IBM has managed to create interest in SAA among users. Since March 17 last year IBM has briefed at least 200,000 US users. In the US, users in the largest companie
s now have a very good understanding of the SAA mission and potential benefits to them. They are probably one year ahead of their European counterparts who are now starting to show an interest in SAA, Killen said. Killen has just written a book to be published in June on how the idea of SAA evolved within IBM. But are users, software vendors or IBM actually developing SAA applications? Languages, application enabling tools and communication software are available from IBM. But as yet, Killen conceded that he had not seen any SAA applications. There are reports by some companies that they have brought out SAA compatible products – meaning applications developed for one of the three SAA hardware environments, which they plan to port across all the hardware environments but Killen admitted he didn’t know of anyone who had really done that. In addition to the generally held view that SAA has not yet taken off is a feeling here that UK users seem to have got a rough deal where SAA standards have already been set. The main complaint centres on the absence of RPG from the list of SAA languages although it is the most widely-used language on System 36 – and RPG III was once the only language on System 38. PL/I, which is popular among UK IBM mainframe users is also absent from the list. And Basic, also popular in the PC-DOS environment will not be supported under SAA. According to Ms Jones, PL/I and RPG were not included in because there are no accepted international standards. RPG is a top priority to be included in SAA. There is no international standard for RPG. IBM has three. Product plans also have to be in place for PS/2, System 3x, VM and TSO before IBM can announce it. We won’t announce something as SAA until we have the made the commitment to provide it in all three environments. With RPG this is extremely likely to happen – but I would not say so of Basic because there is no requirement for it on System 3x and System 370, she said. But users are looking for stronger assurances from IBM. Michael Moore of the small systems group at the IBM Computer Users Association remains concerned despite IBM’s plans for incorporating RPG into SAA. I have seen nothing in writing yet, he said. Killen said in the US IBM has managed to allay users’ fears about the company’s commitment to SAA. In the last few months IBM has told mid range users that it will definitely support RPG under SAA in the relatively near future and that PS/2 will run RPG too. What about PL/I? Once they get round to looking at SAA in detail users may start wondering whether they will have to re-write their PL/I applications in order to comply with SAA. Again IBM sounded reassuring, without wanting to commit itself. I would expect PL/I to be incorporated within SAA. It’s a very high priority to include it. There is an emerging ANSI standard PL/I so users won’t have to adapt their products to Fortran. But Killen suggested that users may have a tough time getting IBM to incorporate PL/I into SAA. To my knowledge there is no plan to support PL/I. I know of no-one in the US that’s pushing for it it’s popular only in IBM’s own development labs and in Europe. Killen added that users have also shown concern over the fact that IBM is committed to its DB2 relational database DB2 but has not made it clear whether IMS will also be supported. Some big users want IBM to support IMS but IBM seems reluctant to do so because few users would want to run IMS-type applications in the mid range and PS/2 environments, he said. Performance problems There are other issues that users will also have to address soon. System 3x users are already concerned that once Silverlake has been announced IBM may put most of its backing behind it, effectively withdrawing support from the current System 3x product line. Users must also decide whether they should develop on PS/2 and execute under another environment or develop under VM and execute in the PS/2 environment. Questions about how easy it will be to port applications across the SAA systems also remain unresolved. And while SAA enables
co-operative processing so that databases and applications on PS/2 can work in conjunction with products on mid-range and mainframes it is not clear yet how well this will work. Killen said users expect performance related problems in this area. Yet SAA remains a side issue for most users. Signs are that most have not yet come to understand what SAA is and its crucial significance to IBM’s future product strategy. Users – particularly in the UK and Europe – need to lobby IBM even more strongly to ensure that their concerns about SAA are pushed to the fore. Failure to do this will result in increasing apathy towards SAA.