Does the computing landscape have to be so fragmented?
When it comes to software, there is always more than one way to do the same thing. Martin Brampton, director at consultancy Black Sheep Research, wonders if life could be made easier...
One of the hardest parts of IT is getting all the bits to work together. Even old, established industries like the railways find it something of a problem. On my regular journeys between York and London I sometimes travel on GNER's White Rose service, using ex-Eurostar trains. They have a quite basic compatibility problem - the trains are longer than most of the platforms.
GNER has resolved the problem with jargon worthy of the IT industry. The trains are equipped with SDO (Selective Door Opening) which means doors on the coaches that might be outside the platforms are locked shut. SDO is not a very sophisticated solution because the doors stay shut even in the stations that have long platforms.
This is very much the kind of thing that happens in IT. Something new comes along but it is not entirely compatible with what is already installed.
Windows is a perfect illustration of this. The early versions were built on top of DOS, a system that was designed for the earliest PC processors. Enabling Windows to evade the increasingly onerous restrictions of DOS was rather like Selective Door Opening. Once you are in the Windows environment, it is more spacious, but getting in and out is awkward. The messy interface has persisted right up to the present, only to be supplanted if Microsoft can persuade everyone to adopt XP in preference to DOS-based systems such as ME, 98 and so on.
Similar things happen at a conceptual level. As computers have emerged from back rooms to be used by almost everybody, so the demands on software have changed. Designing software as a model of the real world has had increasing appeal, encouraging the notion of software objects. Latterly, components have become more fashionable, but the key principles are much the same.
A major stumbling block for objects has been the difficulty of storing them in relational databases. Excellent in their own way, relational databases were created to handle data that could be easily decomposed into simple elements. Objects are not simple elements, leaving a mismatch that is hard to handle.
Linux has been a shot in the arm for server compatibility. While the top Unix vendors were slugging it out at the high end, Linux was busy taking a large slice of the PC-based server market. As a result, there is widespread availability of applications for PC Linux.
Two separate trends are now creating a convergence. Vendors are making their different Unix versions compatible with Linux applications, and Intel's 64-bit architecture is becoming the basis for a broad range of computers. That leads to the possibility of compatibility across many different platforms for pre-compiled Linux applications. This is the critical condition for shrink-wrapped software. Of course, this does nothing for the many incompatibilities between Linux and Windows.
IT managers have to formulate policies to cope with the difficulties and frustrations inherent in incompatible technology. When a guiding principle is sought, there are two main rivals.
For some time, the more popular view was that business processes should drive technology choices. That sounds entirely sensible until one contemplates the resulting mishmash of incompatible systems.
Then, the rival principle of deciding on an IT architecture becomes attractive. Unfortunately, that usually breaks down when someone in the business wants a piece of software that falls outside the chosen architecture.
At one time, software developers could seriously contemplate starting all over again, using the latest ideas and technology. That is now recognised to be impossible, yet there is immense pressure to integrate systems to support business processes such as customer service initiatives. As Oliver Hardy used to say: "This is a fine mess you've got us into." Or something like that.
** Martin Brampton is a director and founder of Black Sheep Research (http://www.black-sheep-research.co.uk
), an independent consultancy providing research, writing and speaking services on a wide range of business and technology subjects. Martin was previously a director at Bloor Research, and has worked with IT as a user and analyst for over 20 years. He is a frequent contributor to silicon.com's weekly Behind the Headlines TV programme and can be contacted at email@example.com.