Windows or Linux ? - A case for logic, instead of emotion.
Following a very interesting Letter to the Editor in the May 2004 issue of the BCS Computer Bulletin entitled "Windows or Linux?", I penned the following reply.
Dear Editor,
I must write to you in response to Stephen Jay's letter ("Windows or Linux?" - The Computer Bulletin May 2004), which appears a call for the use of Logic, rather than Emotion, in making IT decisions. In this, I applaud him.
However, I must draw his attention to a number of inaccuracies.
Firstly, I do not think that there is a belief among IT managers that "the dominant brand must be the best". Most IT managers I deal with are well aware of the so-called the "Betamax effect" - i.e. a superior technology can be out-classed in the marketplace by a less technically advanced one.
Whilst working as a Production Architect, I have sometimes chosen less technically advanced solutions, over superior ones. The reasons were pragmatic. The products chosen had larger market share, and therefore the suppliers were considered to have a better long-term future. Support staff were easier to recruit. Experience was easier to find among the users. There were more Independant Software Vendors using the product, and hence more access to Packaged Solutions. It was not that the dominant brand was "better" - far from it. A balance of technical and non-technical issues had to be struck.
Secondly, I take issue with Mr Jay about his assertion that Linux "was built for server computing". From my reading of Linux history, it was originally built as an academic task. Server issues - multi-processing, scalability etc. were added later, with considerable success. One could argue that NT (built by designers of VAX-VMS) had a more robust server-orientated design, at least in the beginning.
Thirdly, he says that there is "no catch" with Open Source. Wrong again. There are risks with Open Source, mainly relating to support. It is a fact that any changes to GPL code (e.g. bug-fixes) must be re-issued into the GPL community. This acts as a dis-incentive for companies to support code without additional renumeration. This is a risk. It can be obviated by careful choice of product, and looking hard at who will support the product (including the possibility of self-support, with it's associated costs). But it is a risk nevertheless.
I am not arguing for either Windows, or Linux, as a standard solution. As Mr. Jay says, the decision has to be made on "benefit and value".
The word "religion" used by Mr Jay when referring to IT choices, is insightful. Most IT manager are required, from time to time, to take a "leap of faith", to "make the call" on a subject, based on what appears to be inadequate information.
But in order to do that professionally, we must attempt to take into account as many of the facts as possible (both technical and non-technical) and not allow ourselves to be blinded to the risks.
We use our experience and knowledge to assess the issues involved, and then be prepared to back our conclusions.
It's what we are paid for.
Yours Sincerely,
Dennis Adams, CITP, MBCS, BSc.
This letter was subsequently printed in the September 2004 issue of the Computer Bulletin