This site will look much better in a browser that supports web standards, but it is accessible to any browser or Internet device.
Crucial to understanding the development of the PC is the relationship between Microsoft and IBM. By the late 80s this was under some strain as Gates' deal with IBM had allowed MS to produce operating systems for clone manufacturers. With microprocessor technology making rapid advancement, it was apparent that an improved system would be needed to take advantage of these new chips. MS-DOS, however, was Gates' "cash cow". How to keep this in play, without impairing the relationship with IBM, was becoming Gates' chief concern. Visits to Xerox PARC had convinced Gates that a GUI was necessary.
IBM, although the major player in the PC market, was still dissatisfied with sales. It regarded any sales for clone manufacturers as lost sales. IBM developed a strategy to counter this problem. They were convinced that the key to market dominance was the development of a new OS. This would tie IBM customers to their machines and IBM software. Work on a successor to DOS begun in collaboration with MS in 1983. This was followed by a formally agreed joint venture with MS in 1985. The success of OS/2 was not in MS's best interests, but Gates thought it important to keep a relationship with IBM. Work on OS/2 suffered because teams of developers were spread across the country. Furthermore, IBM were insisting that the OS was compatible with their AT boxes, which were based on Intel's 286 chip instead of developing a OS to take advantage of the more powerful 386. Consequently the early version fell between the two stools. The final version did meet critical approval, but at a price. Multitasking was now possible, but OS/2 required more memory, and expense, than the then current PC specifications. It didn't take hold as the de facto OS. Most users decided to stick with DOS rather than part with more cash to buy the extra memory and the new OS.
Gates attempted to interest IBM in Windows, but the latter wanted to develop a proprietary system. In 1989 a compromise was reached with Windows being aimed at low end users and OS/2 at the more powerful machines. This was never going to succeed. Software developers were now faced with deciding which platform to support. Was it to be Windows, OS/2 or DOS? As the relationship with IBM deteriorated Gates, pushed the development of Windows. Early versions were far from successful and Gates underestimated its complexity and hence development time. But success was eventually achieved. Windows was promoted by Gates as the successor to DOS and not OS/2. Windows GUI basically sat on top of DOS and thus was compatible with existing DOS software. Since MS already had immense leverage as the supplier of OSs to most PCs on the market. Although less powerful, it did not have the drawback of requiring more memory. Gates was able to take full advantage of MS's existing sales network. Gates understood the PC industry, IBM did not. Users are extremely reluctant to change their OS. IBM's rivals supplied their machines pre-installed with Windows. By the time OS/2 came to market, in 1992, Windows had already sold several million copies.
Often, software is not developed from scratch. New hardware development is often ignored and the development process is restricted by a perceived need for the software to be backwards compatible. For example, DOS continued to use 640K memory addresses. Legacy problems are particularly acute in the mainframe business where hardware is very expensive and add-ons are continually being produced in order to maintain use of the older technology. These are called legacy systems. This led to the Y2K problem. When mainframe systems were built in the 60s and 70s memory was expensive and so to compensate, engineers restricted dates to 2 digits instead of 4. Huge amounts of money were thus spent trying to resolve the problem.
There are three main functions of an OS.
Sometimes an OS has other functions. If a network of PCs has to simply access printers, then Windows will suffice, but if users need to access servers or other computers on the network then other types of OS are needed. Similarly workstations and minicomputers required more complex systems.
Developed at Bell Labs in 1969 by Ken Thompson as an OS for a minicomputer. Initially written in B language and later developed with C, which is used for current versions. Using a high level language to write an OS was unusual at that time. Chips are programmed with Assembly language, which is close to machine language. Writing an OS in Assembly guaranteed speed. Writing in C, meant that UNIX was relatively easy to port and UNIX acquired support from software developers for whom C was the language of choice. Bell Labs released the source for UNIX ensuring its growth and popularity in education and research centres. UNIX is available in different "flavours", each incorporating a standard general structure, but individually developed for specific functions.
UNIX is developed by small teams writing small components, which are then combined to to create the OS. The core of the OS is called the kernel and contains the minimum number of instructions that the system needs. The kernel is quite small and is in keeping with the overall UNIX philosophy of "small is beautiful". It is portable across many architectures and although PC versions are available there is some user resistance to it. UNIX supportrs enjoy its command interface, claiming it is faster and allows greater control than a GUI. It was one of the first OS to allow multi-tasking.
The computer industry is marked by its intensive competition and collaboration, sometimes involving unlikely participants.
This all leads to constant change in the industry.
Next page » About the Internet section
Previous page « Apple and the Mac