IntelliCorp Inc, headquartered in Mountain View, California is destined to be in the news frequently this year as a deal between it and AD/Cycle partner KnowledgeWare Inc comes to fruition. Although the negotiations between the two have engendered much gossip, a lot of industry watchers are puzzled as to why KnowledgeWare needs to tie up with a Knowledge Base Systems company, such as IntelliCorp. However, IntelliCorp has a very advanced software engineering environment to offer KnowledgeWare – it has expertise in both object-oriented programming and in the Unix and Windows 3.0 environments, all skills that are at a premium these days, it would seem, if you are a CASE vendor attempting to work with IBM.

DNA structures

The problem many people have in perceiving IntelliCorp’s value is that they tend to view it as an expert systems company with a penchant for Lisp. A while ago this was a true evaluation of the company that was founded in 1980 by four Stanford University scientists. Indeed, the company, then called IntelliGenetics, was born out of a genetic engineering project as these scientists used Lisp to model DNA structures. The software modelling core was soon developed into a general purpose modelling environment and the four realised they had a generic product on their hands – the Knowledge Engineering Environment, known as KEE. After an initial public offering in 1983, the company changed its name to IntelliCorp to reflect its product’s broader market appeal and retained IntelliGenetics as a subsidiary company. Marketing deals with Sperry Corp, Hewlett-Packard Co and Texas Instruments Inc followed swiftly. International headquarters were established in Germany in 1985, with a UK subsidiary being set up the following year. Then followed what UK managing director Russell Prince-Wright calls the AI winter of 1987 when no one wanted to buy into anything let alone a relatively new technology. This was followed by something of a boom period in 1988 when expert system tools began to take off. However, the period of retrenchment focused IntelliCorp’s mind on the Lisp market and its future direction. IntelliCorp knew that it wanted to provide a rich software development environment written in a commercially marketable language. It watched Inference Corp attempting to do the same thing with its Automated Reasoning Tool product and failing to satisfactorily bridge the old and new products, though Prince-Wright respects both Automated Reasoning Tool and ART/IM – he simply questions the attempt to bridge the two products with C-ART. Consequently, IntelliCorp says that it decided to create a product from scratch – a dif-ferent product environment for the Unix workstation written in C that was capable of delivering software to a variety of environments. Furthermore, IntelliCorp was committed to Lisp for a long time and didn’t want to outdate the KEE Lisp product range as it serves its purpose for the AI specialist.

By Katy Ring

At the moment the KEE market has stabilised and has grown slightly since the release of version 4.0, which has a colour X Window implementation. Indeed, any technical person employed by IntelliCorp will sooner or later reveal that Lisp is the best programming language in the world but that it missed the commercial boat and now KEE is des-ined to be popular in a small niche market. Understandably, the corporate decision was taken that if IntelliCorp wanted to grow into a software force to be reckoned with it would have to move with the mar-et. And so it was that the modelling capabilities of KEE were taken and developed in a C environment into a full mainstream life cycle development system that is now cal-ed ProKappa. ProKappa offers a methodology for an object-oriented rule-based approach to programming and has three engines: an object engine to manage objects, a rule engine to manage rules, and a data access engine to manage the mapping of data from databases onto objects. One of the neat things about ProKappa is that it has good dynamic capabilities through its Runtime Object Manager that enables

any changes that might be made at the development stage to be made at the runtime execution of the application. Furthermore, the end-user is able to create and delete classes and instances at runtime which adds up to a flexible development environment offering a developer’s graphical interface, the option to code in C or in IntelliCorp’s ProTalk, which is a kind of pidgin English and is integrated so that applications coded in ProTalk are compiled in C, incremental loading and linking and a data access tool that stores and retrieves ProKappa data in relational databases using Techgnosis’ SequeLink. While this development was going on IntelliCorp had decided to have an object-oriented programming environment for Windows developers. At the 1989 American Association of Artificial Intelligence trade show the IntelliCorp clique watched MegaKnowledge demonstrate its Kappa PC product and were so impressed they bought the company. The code and all the employees came over to IntelliCorp and 90% of the code was re-engineered to make it fit with IntelliCorp architectures – code from Kappa PC is upwardly compatible with the ProKappa environment.

Fun to use

The Kappa PC environment offers an object browser, knowledge tools, an application language KAL that compiles down to C, a session window, a rule relations window, a rule trace window and an inference browser. The product has direct dynamic data exchange links into dBase and Lotus 1-2-3. ProKappa is aimed at the engineering, finance, defence, communications, scheduling and management sectors, while Kappa PC has built up a following in the insurance, building society and telecommunications markets. Kappa PC was launched a year ago and in its first six months 4,000 copies were shipped. IntelliCorp likes to describe its Kappa products as being standards-based. In terms of Kappa PC this simply means conformance to Windows 3.0 specifications. However, IntelliCorp was one of the early members of the Object Management Group and is involved in the Task Force working on the definition of the Object Model. The company believes that ProKappa will be one of the first software development environments to comply with this standard when it is set in November. It is also an environment that is easy and fun to use and it seems to explain why KnowledgeWare should be interested in a Knowledge Base System company. For IntelliCorp’s part,it has seen financial performance slide, but it has $13m in the bank and no long term debt. However it wants to grow by moving into the mainstream software development market…