America’s flagship pre-competitive research co-operative returns to its roots
The Microelectronics & Computer Technology Corp is going back to its consortial roots, a representative of the pre-competitive research consortium said. Set up in 1983 at the initiative of William Norris then chairman of Control Data Corp, in partnership with most of the major US computer and semiconductor companies bar IBM Corp, in an attempt to enable the US to narrow the gap with then all-powerful Japan Inc (CI No 19) it has maintained a low profile since the beginning of the 1990s. The Austin, Texas-based consortium has undergone a few changes in that time, but is now back on its original track. In 1990 it appointed a new chairman and under his direction the co-operative decided that it had good technology, but it was not creating sufficient financial return for its members. The consortium concentrated on developing spin-off companies to gain something back from its investment in research. However, it began to feel that too much emphasis became placed on the process of spinning off companies and not enough on underlying research. It has now reverted to its original philosophy and is re-emphasising collaborative research. Although MCC still wants to pursue spin-offs as appropriate, it doesn’t want to do it at the expense of neglecting consortium projects. Another recent change as a result of the new shift is that MCC Ventures Inc, the venture capital arm of the consortium that was set up in September 1992 has not been active since September of this year. MCC said that this was because it gave the impression that it was shifting to becoming a venture capital organisation. There are currently four spin-offs from MCC, all based in Austin. The first to be created was Evolutionary Technologies Inc in January 1991 with its data migration tool suite for automating the migration of data between dissimilar data storage environments, followed by Corporate Memory Systems Inc which provides the CM/1 groupware tool to combines a graphical hypertext interface with an object-oriented database and enables users to construct maps of important communications and information relating to a project such as ideas, questions and external documents. In January the next year Pavillion Technolgies Inc set out to commercialise Process Insights, an adaptive process control software and in July of the same year Tamarack Storage Devices was spun off and set out to commercialise the consortium’s holographic storage technology.
Cyc Common Sense Knowledge Base contains the information taken for granted by humans that leaves computers stumped
The Cyc Common Sense Knowledge Base project at the Microelectronics & Computer Technology Corp has been working for the past 10 years to build up an encyclopaedic knowledge base of information that is generally considered to be common sense. The project could be described as filling in the white space of an encyclopaedia. It effectively contains the knowledge that the writer takes for granted and contains two million rules including time and spatial relationships. So, for example, in an explanation of the heart, an encyclopaedia definition might assume the reader already knows that the heart cannot be seen outside the body and it is this kind of information that the Cyc knowledge base contains. The idea behind the project is that expert systems generally become limited when you get to the edges of their domain of knowledge, so that a general database of common sense knowledge could help improve these systems. At the end of this year or the beginning of next, Cyc Corp will be created to work with MCC for the commercialisation of applications. Applications of an experimental nature have already been put together, such as data checking in spreadsheets and decision support, and applications like these will be expanded commercially. Microsoft Corp is one of the companies investing in this project.
Carnot project addresses management of disparate databases
The Carnot project is working on the development of a set of software tools to e
nable databases within different organisations to be managed across distributed networks regardless of whether the systems are compatible or not. The tools developed are written in C. The idea is that the databases can be managed and queries executed on them from different databases that are not compatible. So different systems in different departments or even different organisations altogether would be able to communicate effectively with each other. This Microelectronics & Computer Technology Corp project concludes at the end of the month and two new projects will then begin based on the research already carried out by the Carnot project. The new Crossflow project will use distributed systems for supporting business process changes, while Infosleuth will work to establish intelligent agents for datasearching across distributed networks. A small Austin-based real estate firm, Express Star Services, has been using the system commercially since last month and is working with accountancy firms Anderson Consulting, Ernst & Young and Price Waterhouse & Co to apply it to property.
Content from our partners
Mass data
The companies are developing an information integration system to support property management investment where multiple buildings have to be managed and a variety of data held on each one, such as who is occupying the building, details on reports that have to go to investors and regular marketing that has to be sent out. One of the Baby Bells is experimenting with the Carnot software and working on developing an application in at least one of the regional offices to manage and speed up its service provisioning. When a customer applies for a new telephone system, a whole range of factors need to be accessed such as customer credit history and billing files need to be set up. The service can be lengthy because of the many databases that have to be accessed, but with the simultaneous database access, the service time could be cut from around two weeks to something like one day. This information resides in different databases and in the case of credit histories, possibly in different organisations. Effectively the system is a means of sending out and receiving information across different and incompatible systems and has a niche now client-server systems are emerging. It also promises to enable easier access to mass data even if the data is held in legacy systems. – Abigail Waraker
