The Eclipse conundrum

Comment Eclipse, as a development platform, is taking the Java development world by storm and is likely to prove the IDE (integrated development environment) of choice for that community for some time to come or, at least, until something better comes along.

However, Eclipse provides a conundrum for CIOs and managers of IT development teams. The reason for this is that Eclipse guarantees interoperability between the different plug-ins that are available, and provides a common (up to a point) look and feel. Let's take this another step: Eclipse means that you can take any of, say, 15 different coding environments and plug them into Eclipse and they will all work not just with the other development tools that you use, but also with all of the other coding environments.

In other words, it would be perfectly feasible to have a development team, each of whom used a different coding environment. Indeed, there would be nothing to stop any individual using multiple coding environments, perhaps because he or she preferred this tool for web applications, and that one for building web services, and a third for something else.

Now consider this proliferation from a management perspective. Is this a dream or a nightmare?

Before we even attempt to think about this, there are some constraining factors we should consider. For example, you could just as easily have multiple requirements management products or multiple software configuration management solutions. However, if you did this, then how would you keep track of which tools were being used to manage which projects? You would need another super-management tool to manage the management tools – and, of course, you could have multiples of these - which would mean that you would need super-super-management - and so on - which is clearly daft. So, we must logically have some control over which management tools we use.

Another consideration is one of support. If you want to have formal support from a supplier, together with things like indemnification, then you are obviously going to need to limit the tools you use for cost reasons, though this does not, of course, prevent you from using other, unsupported software.

A possible counter-consideration is training. One can imagine a new development team being put together, in a few years time, made up of developers from a variety of different backgrounds. Each of these people is likely to be familiar with a particular toolset - why bother to retrain them, with all the costs and time involved in that process – why not just let them use what they want to?

On the other hand, there is another downside to the open-for-everything approach: developers like to play with the latest tools, and job adverts require up-to-date skills, so there is a danger that developers will use any open-handed policy more to increase their value in the job market than to do the job in hand.

So, you can see the issues involved and you can probably think of even further ramifications (it certainly has some for vendors). Unfortunately, I cannot say that I have any solutions to recommend and in any case, they will probably vary from company to company. But a clear Eclipse (and, indeed, Open Source) policy is going to be needed by any organisation pursuing this development route.

Related stories