Screening blue murder

The three-finger salute of ‘ctrl, alt, delete’ has become second nature to most users. But engineers need a stable computing environment, so why do we put up with Windows? asks Charles Clarke.

Very rarely do systems selectors consider all the options before choosing an operating system. All too often it is selected for non-technical reasons, corporate policy or just plain bias. There used to be a financial incentive for choosing Windows over other systems, but an OS that leaves designers cursing the ‘blue screen of death’ on a regular basis neither optimises nor maximises a business’s most valuable asset.

The OS is a critical component of any IT solution, but all too often it is seen as a ‘decision driver’ rather than as the consequence of any rational decision-making process. Systems are enabling technologies to satisfy the requirements of the preferred solution.

That solution has to deliver the appropriate applications to the user in the best way possible. Any approach where the OS is conditioned by word processing and accounting systems has to be questioned. Many businesses cite standardisation, but there’s no evidence that this is areasonable basis to drive a decision.

Indeed, the success of the Love Bug virus and more recently the MyDoom andNetSky worms is an example of thedangers of standardisation. Many people blamed Microsoft for the damage that these intrusions caused. But is Microsoft really the problem? Its software does exactly what it says on the box, so maybe the fault lies with the huge numbers of people who bought it.

System choice is an emotive subject which often defies rational thinking. The goal of any well-considered IT solution is to pay for what you want rather thanletting your budget dictate what you get.

Too many decisions are taken where the price is imposed through budgets and all the rest is expected to follow.The four main factions – devotees of Windows, Linux, UNIX and Mac – are set in their views. Novice users need to be cautious of any OS recommendations as the rivalry runs deep.The traditional MCAD vendors seized upon Windows in the mid-1990s when there was a significant price differential between UNIX and Windows. Vendors were offering solutions to companies with fixed capital budgets, and the ‘low-cost’ Microsoft platform enabled them to increase their percentage of that budget.

For vendors who offer both platforms the cost of developing their Windows solution far exceeds the costs of developing all of their UNIX versions. And this is without the added costs of support that arises from the adoption of the Windows model. Would they, for instance, have gone for Windows if Linux had been viable in the mid-1990s? Microsoft makes money from the process of change, not the delivery of stability.

Bill Gates has already announced that over the next few years he intends to fundamentally re-engineer Windows 2000 and XP Longhorn, the ‘completely new’ version of Windows originally due in 2005. but that date seems to be’slipping’ already – another Microsoft trait. Is this consistent with the need to maximise the value of your most expensive asset? Just imagine the bills for new hardware (old kit rarely runs new versions of Windows well) and the re-training. And should Windows products be at the heart of any enterprise when it spends years telling us that NT4 is an ‘enterprise-capable’ operating system, and barely 12 months later releases Windows 2000, making NT4 obsolete?

Is it time to re-appraise the UNIX/Open Source situation? Traditionally UNIX is seen as ‘premium priced’, although this is changing fast. UNIX offers a stable platform where there is no conflict between hardware and software. One company is responsible for the integration of everything and for making it work. Another often-overlooked benefit of UNIX is that it is free of unnecessary frills.

The system is just left to do its job, with users coming in, logging on, doing their work, logging off and going home. There is no reconfiguring of desktops, installing new wallpaper or worse, loading applications that come ‘free’ with a CD in a magazine. Linux is not just another flavour of UNIX. Offering many of the attractive features of traditional UNIX, it also comes with two of the major disadvantages of Windows. It is still trying to deal with a complex multi-vendor hardware model and is very accessible to hackers and viruswriters.

While not as vulnerableas Windows, its Open Source nature is not consistent with total security.Windows has enjoyed a fairly chequered reputation for its robustness, but we have all come to tolerate system lock-ups, slowdowns, memory leakage and viruses, and the three-finger salute of Ctrl, Alt, Delete has become second nature to most users.

Engineers need ultra-fast machines that work all the time. So why do we put up with Windows?

The answer to that is maybe in critical areas we don’t. Take the funky new Williams F1 car. I have it on good authority that the CFD for its quirky nose was done on a Linux cluster.