Oracle the CASE vendor moves out from the shadow of the database to flex its muscles
It is a not much publicised fact that Oracle’s CASE product and consultancy business is worth around $100m – big enough to be a stand-alone software business in its own right. And it is a home-grown UK operation fostered by Geoff Squire and Richard Barker that began with an internal tool called the SQL Design Dictionary sitting on top of the Oracle database in the mid-1980s. Customers started asking for the tool and four years ago Ellison decided the time was ripe to move into the CASE market – the necessary technology resided in the UK company and that is where it has remained. Being a database, company Oracle took the working from the database backward approach to creating a CASE toolset. The whole concept is described by Oracle consultant Cliff Longman as the database for developers approach to software engineering. Because of the origins of CASE*Dictionary, Oracle’s CASE strategy has historically been tied to their repository – for example, their entity models from CASE*Designer are automatically checked against the dictionary. However, although in many senses this leads to a very closed toolset, it has the advantage of offering true multi-user access across a network because the dictionary was developed to run on a Unix server. According to Ian Fisher, Oracle’s worldwide director of CASE marketing, most of the rival Information Engineering products were designed for mainframe use and although data can be downloaded from the mainframe to the personal computer, this leads to islands of development with concomitant integrity problems. Nevertheless, nowadays closed repositories are becoming a little suspect, which is why Oracle is developing a close relationship with Bourne End, Buckinghamshire CASE integration company Software One. Oracle will be marketing that company’s Exchange product as CASE*Exchange thus enabling Oracle tools to plug into a variety of repositories – see below. The driving force behind this open repository thinking is, ironically, the need to plug into IBM Corp’s Repository when it arrives. To this end it is Oracle’s intention to join the IBM Vendor Assistance programme, which means that it will get quarterly AD/Cycle updates, although Oracle is clear that AD/Cycle represents no software engineering solution for non-MVS users. At present Oracle has around 5,000 installations of its CASE*Dictionary and 80% of these are going to VMS, Unix and OS/2 sites – indeed, Oracle shipped 1,200 copies of its CASE tools on OS/2 in the US in the six months to the end of March. The remaining 20% of installations are in MVS, VM and MS-DOS environments. In fact because of this deal Oracle’s CASE products are beginning to be capable of reaching the parts that the Oracle database cannot. For example, there is some internal conjecture that it may now make sense to put Oracle’s CASE tools up on the AS/400, where it clearly does not make sense to put the database. The growing relationship with Software One also means that the two companies will collaborate closely on future developments. Oracle has mentioned future products such as CASE*Bridge and CASE*Connect before, the former being a bi-directional tool to bring non-Oracle repository objects to and from CASE*Designer, the latter enabling CASE*Designer to pull information out of DB2. Both these products are now likely to be developed in partnership with Software One. Further down the road Oracle is developing CASE*Project, a project manager workbench that enables the project manager to manage the different pieces of the life cycle via the repository. In all, the Oracle CASE product set is looking in good shape – it has been the recipient of a huge research and development budget over the past 18 months to two years, which has enabled the developers to create a flexible, layered architecture model underneath the tools making them highly configurable for co-existence with other vendors’ software as well as enabling users to tailor the tools to suit themselves. It is the most signifi
cant CASE player around today to come from the multi-user mid-range hardware environ-ment and it is strongly dictionary-oriented. However, Oracle CASE has been operating under two major handicaps for the past couple of years – it has been physically tied to the Oracle database and it has been seen internally as a poor second cousin to the database. But both these handicaps are now being removed via the deal with Software One and a changing internal culture – after all the rise of Geoff Squire within the company will undoubtedly have raised the profile of the UK operation and its research and development strength in CASE.
Software One continues to grow and will take its products, strategy to US
A year on (CI No 1,511) Software One Ltd, Bourne End, Buckinghamshire is still a firm that bubbles over with enthusiasm, a quality now tempered with a realism developed over the year or so spent operating in the commercial arena. Michael Hunt was brought in as chief executive a few months ago to improve Software One’s credibility – Hunt was, prior to the appointment, executive vice-president of international operations at Dun & Bradstreet Software Inc – and get it speaking to users and vendors at board level. The next 12 months will see expansion in the US with an office in Atlanta, Georgia, followed by growth on the continent. Fresh equity funding is being negotiated to help fund growth, both organic and acquisitive. Two new releases of Exchange will arrive over the next six months – version 1.7 will offer a meta language giving users greater control of the data passing to and from Exchange. In the New Year version 2.0 will arrive enabling data and processes to be passed via Exchange, which will by then operate on Unix and OS/2 as well as MS-DOS, where it now resides.
Software One’s Exchange offers a de facto standard for CASE integration
In developing Exchange, Software One has worked from the premise that to integrate CASE tools properly, a generic repository is required, independent of the conceptual architecture to be found in analysis tools and from the design architecture to be found in design tools. When each proprietary vendor attempts to glue its tools to another vendor’s tools, the fix tends to end up corrupting one architecture or the other. Software One co-founder Andrew Mercer explains that Exchange offers a logical architecture between the two where source meta data is converted to Exchange standard form, conceptual data objects are resolved and design level objects are inferred from logical level objects. Software One is in a unique position to do this: it is independent and so can work at source code level to integrate proprietary CASE tools of participating vendors. Indeed, it is rapidly turning into a de facto standardisation operation, providing much tighter integration than can be found at the Open Software Foundation, Object Management Group or CASE Data Interchange Format body. – Katy Ring
