PLUG AND PLAY SOFTWARE

by BERNARD A HODSON

In "plug and play" hardware one links different peripherals seamlessly to the computer, the software in the computer recognising the new presence and arranging to integrate it with existing hardware attachments. Plug and play software is never likely to reach that stage with today's approach to software but some progress is being made, one step being IBM's "on demand" marketing strategy. The concept, however, has been around for years, and is just lacking a sea-change in software concepts to make it happen...

What is meant by "on demand"? In theory it is like "just in time" inventory so loved by manufacturers. The manufacturer does not stock items needed to build a product but relies on a parts vendor to deliver needed items at the time they are needed in the manufacturing process. It offers significant savings in inventory costs.

In "on demand" software operations a user does not buy software packages but has the application software transmitted when needed, paying a rental cost. The problem with current "on demand" software is that applications are large and require sizeable bandwidths. It is not "plug and play".

"Plug and play" application software will soon be on the market and pseudo "plug and play" operating systems can also be developed. The developing "plug and play" techniques for applications, however, will not need an operating system in their final configuration...

While the elimination of the operating system is on the horizon, it will be with us for a few years, so what could be done there. An examination of one of the Windows products shows it taking 598,699,867 bytes of hard drive before a single application can be run. These bytes are divided into 3,862 files of varying size. When you consider that most users need less than five percent of what is provided it represents an obscene resource waste.

The control of such a huge code amount is difficult, involving the establishment of several "interface layers" in a hierarchical system. It is impossible to guarantee that so much code is free of errors, and there are many places where security can be breached, as the industry has found to its $46B annual cost, from worms, viruses, hackers, identity thieves, spammers and the like.

The cable television industry reluctantly developed the technology enabling viewers to pick the shows they want to see, although this process still has a long way to go. Some offer a basic package and the viewer can add other shows to it. There is absolutely no reason why the operating systems vendors cannot do something similar, offering a basic package and then having a user add features of interest...

ComputerWorld Canada, in December 1995, reported that Microsoft were about to change programming and ease the legacy burden by developing what was known as Intentional Programming (IP), involving the construction of a large number of reusable elements. Microsoft offered it as something new but the technology to do IP had been developed at the University of Manitoba about thirty years before. IP enabled language mixes but was operating system oriented. The Manitoba development eliminated conventional languages altogether, along with the need for operating systems, while still being able to do such things as IP. A few years later Robin Bloor, a UK guru, showed how the concepts of IP could be generated with the Manitoba pioneering.. It was a flawed step by Microsoft towards "plug and play" software, but was not implemented by them...

You cannot develop "plug and play" software with the current software fragmentation, caused by our adherence to the von Neumann philosophy of computing. The Turing approach, which has been proven successful, is ideal for "plug and play". This philosophy involves the creation of "states of mind" (essentially reusable modules, so IBM is on the right track but heading in the wrong direction) which can be plugged together to build any application desired. Recent developments of the Turing concept have gone even further, replacing the "reusable code modules" by a numeric structure. The numeric approach can eliminate most of the virus, hacker and worm intrusions plaguing the industry and would, if used, counter any attempted IT9/11.