• silicon.com
• Technology
• Software

By Quocirca, 23 December 2004 13:10

COMMENT

When it comes to buying software platforms and applications, businesses don't care about the philosophical differences between open source and proprietary development. They don't even care much about cost. Quocirca's Jon Collins, inspired by input from fellow analyst Dale Vile, explains what user organisations do care about - and why both sides of the debate must accept this to succeed.

The Christmas season is a fine time to remind ourselves there are three topics that should never be discussed at dinner parties, namely politics, religion and sport.

In gatherings of technological types there is a fourth, namely open source and all that goes with it; perhaps because it incorporates elements of all the others.

Equally however, the issue should not be shirked. For the best chances of success, any debate should start and end with the needs of, and the benefits to, user organisations.

Much of Quocirca's research is based on the feedback we get from mainstream IT departments and as far as they are concerned, open source is simply an alternative way for their suppliers to develop software. Once they have the functionality they need, what these people really care about (with the exception of a few hard core zealots) is the quality of the end result, its fitness for purpose, how easy it is to implement and maintain and how well it is supported.

The resulting software, we are told, needs to be delivered and supported by a company that is perceived as a 'safe pair of hands', that is the likes of IBM, HP, Microsoft or (possibly) Red Hat or Novell.

From the enterprise perspective, the evidence we have from both our own research and others suggests there is negligible cost benefit associated with open source products for enterprise delivery. In fact, it is quite common for organisations to tell us that open source actually costs them more from a skills and management overhead perspective. Like it or not, Microsoft's Get the Facts campaign makes a valid point as far as the enterprise goes.

However, and this is a big however, the debate is as much about which applications are available as it is about the cost of ownership. In the data centre most apps in common use are increasingly supported on both open source and proprietary platforms, so the discussion centres largely on support costs. Elsewhere - on departmental servers or in smaller organisations - the existence of suitable applications becomes the defining characteristic.

Once this has been determined, any of a number of factors can skew the decision on the choice of server software, not least of which are:

• Presence of local support, or at least local competence
• Whether there is any money to pay for software licenses
• Minimising the budgetary cost of the server environment
• Time to capability - how quickly can a solution be up and running?

Given these factors, it is perhaps no surprise that the popularity of open source on the server is indicated to be inversely proportional to the size of the organisation. Equally, the factors go some way to explaining the increasing role of appliances running open source software. While Microsoft is helping to make Windows an increasingly attractive appliance platform, many appliance vendors still see Linux as more configurable, cheaper and enabling a smaller footprint than the Microsoft equivalent.

Meanwhile, there is the desktop. Despite wins in certain organisations and geographies, open source has failed to become in any shape or form the de facto standard for the user-facing computer environment. Once again this is largely down to applications. Many common apps still don't run on Linux, so a decision to run Linux on the desktop is a brave one - even if you're up to the job of administering it (better to join the Mac community, where at least you get MS Office with your Unix).

No doubt the discussions will rage for some time but perhaps the time is right for the open source movement to recognise that the debate has moved on. Open source is a methodology, not a religion. If business benefit is the goal, it is more important to focus on what is being developed than how it is being produced. There is no golden rule that an open source product is inherently going to be better than something developed by a commercial software vendor. We should not forget - if only we could all just admit it - that, from a technical perspective, Linux is Unix.

Open source needs to grow up and start delivering fully fledged applications of its own - taking a small but laudable lead from Mozilla, for example, by delivering new functionality (in that case, web browsers that support RSS feeds) rather than reinventing the wheel. There are a great many, potentially very exciting open source projects that are falling by the wayside due to lack of resources, and meanwhile most community effort is spent on developing software that exists already. It's worth asking the question, who exactly does this benefit?

Evolution is about diversification, not specialisation, and we are as unlikely to see a fully open source world as we are to see one that is fully proprietary. While Microsoft and other vendors will continue to fight open source products, it has long since recognised that there is no point in trying to 'win'. The open source movement would do well to recognise this, and should move on as well.