Last week at the SSADM Users Spring Conference, the UK’s Central Computer & Telecommunications Agency launched version 4 of SSADM, the methodology with which software has to comply for civil government information service contracts. As mentioned at the time (CI No 1,423), project mangers will not be over the moon about this new version as it still does not embrace management control or strategic planning techniques. According to Ian Thomson, head of the information systems engineering division for the Agency, this is because SSADM cannot support processes that aren’t linear, therefore as a methodology it cannot contain its own management. Furthermore, while SSADM could could have been extended into the realm of strategic planning, this hasn’t been done as there are more than 70 techniques that could be used for strategic planning and while these could be reduced to an eight or nine part generic model, this would not have been especially suited to data flow diagrams. There are five modules in the core of SSADM v.4: feasibility study, requirement analysis, requirement specification, logical system specification and physical design. As yet there is no guide to prototyping and estimating for v.4, but Thomson hopes they will appear soon – he also encouraged suppliers and users to offer guides on these and other subjects such as risk assessment and quality method. Since v.3 came out in 1986 the market context in which SSADM operates has changed. There are now not only active standards bodies in the US and Japan but the Single European Market is looming and each country still has its own methodology – the UK’s SSADM, France’s Merise and Italy’s Dafne, for example. Then, of course, there is Euromethod, which started out as a European Community exercise to take the best features from each country’s methodology and merge them into one new standard. Nowadays its ambitions are far humbler: to provide a standard framework within which each country’s approach can operate without losing its separate national identity. Anyway the UK Agency is active in Brussels and says it will make sure that SSADM is compatible with Euromethod.

IBM’s AD/Cycle

At present Euromethod merely stipulates that software activity should focus on a standard input and output for programs rather than on standardising the development processes themselves, other than to make sure that each activity bolts on to the one before and the one after. (Vendors will have mixed feelings at the news that the Agency has given the glossary for v.4 more meat for procurement specifications). Finally, there are developments such as IBM’s AD/Cycle to be considered and Thomson says that SSADM will be evolved to fit into the project management side of AD/Cycle and the Repository. As for the back end folk that actually have to write the software, they will probably be fairly happy with v.4. For the Agency claims that a competent v.3 analyst can rapidly assimilate v.4 as the underlying approach is the same for both. The new version is described by the Agency as a broad, imaginative analysis which is less prescriptive than v.3. Furthermore, after carrying out a feasibility study the Agency believes that analysts would probably be better off choosing a non-SSADM method to implement, say, expert systems. However, where SSADM is appropriate, analysts will find that v.4 benefits from the adoption of Jackson-like methods in entity event modelling which means that they no longer have to flounder with data flow diagrams to get from the logical processing stage to the physical design process. Instead, this new version of SSADM has an object-oriented flavour in that the entity life history combines the data with the process, while the effect correspondence diagram means that programs are written for events. These are new techniques to SSADM which should lead to a more methodical system analysis. Copies of the SSADM v.4 manual will go to the National Computing Centre for publication in June. – Katy Ring