Addressing the third Sapiens User Conference was Bob Libutti, programming systems director of market strategy with IBM. The reason Libutti had been invited to address the conference was, presumably, to reinforce the fact that since August IBM Corp and Sapiens Ltd have had a marketing agreement to sell the Sapiens product in the US and Sapiens is now an IBM Business Partner. Libutti started by saying that the Business Partnership programme is a way of doing business that is not totally comfortable for IBM. He, personally has had to do a lot of missionary work with the technical, development and marketing people at IBM to get the programme going. However, he said that once IBM is involved, the analogy should be with a marriage as the involvement in a business partnership is for life. He admitted that IBM’s attitude to Systems Application Architecture has changed since its inception when the architecture was seen as an end in itself. Nowadays, said Libutti, IBM recognises more clearly that Systems Application Architecture is a set of building blocks that it has to get control of. In the user interface arena it needs to define and evolve a set of standards and it also says it needs to control the communications between systems.

Simply by using Cobol you are partaking in AD/Cycle

The message is that Systems Application Architecture is very very much alive – IBM claims that it has shipped 87% of what it promised in 1987. The only area in which IBM thinks it has fallen behind is with its promises for OS/2 because it has been waiting for the 32-bit version. However, IBM did not stand still and today, claims Libutti, offers twice the content of Systems Application Architecture that it promised. However, despite this Libutti admits that IBM has not done as good a job in merchandising what it has done as it might. He stresses that waiting for the Repository is not what AD/Cycle is all about, arguing that simply by using Cobol you are partaking in AD/Cycle. Libutti evidently thinks that IBM is being falsely chastised for the shortcomings of AD/Cycle. He explains that IBM could have launched an ultimate strategy, provided a set of tools linked together and made a big impact. But instead it chose a different path, it chose the flexibility of selecting companies and products in that paradigm for the benefit of the user. In pursuing this strategy, he admits that IBM made some mistakes and was forced to do some fundamental thinking about products that it had never done before.

The idea that IBM’s Systems Application Architecture and related technologies have gone awry is a misconception: almost everything is now in place and everything is on plan. Katy Ring hears the Gospel according to IBM.

Meanwhile IBM has moved rapidly into the Unix environment, which it had not embraced for a very long time because of paranoia in some parts of the company. Now says Libutti, IBM is investing heavily in Unix and is looking closely at the relationship between Unix and Systems Application Architecture. He added that IBM will not make Unix part of SAA, but is working on ways to to get co-existence and co-operation between the two systems: AD/Cycle will be extended to include AIX. That is, a software engineering environment consistent with AD/Cycle with common system functions, common tools that are supportive of the AIX environment will appear courtesy of IBM. Common between AIX and AD/Cycle will be the Information Model with an AIX database and repository being developed by IBM. There will also be increased commonality between Common User Access and Motif. Indeed, Common User Access will also have some commonality with the Apple Computer Inc Mac Finder as Libutti admitted that Apple will have some influence on future versions of Common User Access. IBM and Apple will be working together to develop a user interface that can cope with complex data structures necessary for multimedia applications such as image and audio. In addition IBM is doing some basic research in the area of maintenance in its efforts to produce a sub-framework within AD/Cycle

for redevelopment. This will cover both forward and reverse engineering. Moving on to the Repository, Libutti said that this is a good news, bad news story. The bad news is that the Repository is behind schedules, the good news is that this is because of the necessity of smoothing out the relations between vendors and the Repository will be a better product because of this – the first implementation of the Information Model and the Repository will go ahead in 1992.

IBM is making a major investment in object-oriented technology

As for the Information Warehouse, Libutti quipped that IBM does not need to distribute data as God has done that for us already, instead we need to manage that data. And that by virtue of the Repository, Information Model and Distributed Relational Database Arachitecture is what the Warehouse is designed to do. As far as object-oriented technology goes, Libutti said that IBM was considering using Sapiens tools internally in several industry areas. And prior to its discussions with Sapiens it has been talking to DigiTalk. Indeed, SmallTalk is used extensively in IBM – it was used to build the new Cross System Product, CSP, and was also used to by Intersolv Inc to build its product line for OS/2. Libutti said that IBM is making a major investment in object-oriented technology. This is all well and good except that IBM’s Repository is based on the Entity-Relationship model and not the object-oriented model. However, Libutti says that IBM is extending the Repository to cope with object types, although when pressed said he didn’t know when these extensions will appear. And anyway that still leaves IBM with an underlying database – DB2 – that is unashamedly relational. All that Libutti would say on the matter is that DB2 is durable.