A practical approach to Open Source (part 4)

So far, these posts have covered user benefits of OSS, risks and the need to understand the various different OSS business models.

This final post in the series looks at tips for getting benefit from OSS implementations.

The first point to remember is that even though OSS software is ‘free’. its adoption needs to be planned carefully. This is even true forthe lowest risk area of personal productivity tools, surprisingly. Most people think this is the easiest OSS area, and are happy to let OpenOffice or Firefox or any of the multiplicity of user tools be loaded….but this can lead to trouble. Remember that one characteristic of OSS is that because it is free, there is no need for a user to gain purchasing approval, and hence adoption can be uncontrolled. If it then turns out that the new software introduces an incompatibility with other systems, this can be a nightmare. Make sure that adoption is controlled and that every employee is aware that OSS software purchasing should go through the same governance procedures as commercial software, even if it doesn;t cost anything.

This leads to the next point. In fact, OSS software is never free. The LICENSE may be free, but there are users to train, developers to educate, support to be arranged, risks to be evaluated and countless other tasks. Indeed, at the OSS infrastructure level as opposed to end-user productivity tools area, most OSS offerings are of the ‘framework’ type where the user is left having to do extensive development and customization before the software can be used productively. So, to succeed with OSS it is necessary to evaluate the business case taking all these resource and service requirements into account, even if license costs are 0.

The next tip is think ahead. Ask yourself why the software is free. Is it because the software community, out of the goodness of its heart (!), wants to share its bounties with everyone for free, or is there some other game plan at work? A number of OSS projects have come about from users wanting to find a way to defray their costs of supporting their home-developed code base. Projects such as AMQP and Swordfish srping to mind. The issue here is that if the particular project never really gets popular acceptance, then  future updates are at risk of dying off if the original authoring company changes direction. Other OSS projects are offered by vendors having a ‘full-function’, priced version of the software. Remember IONA’s Artix / Celtix ESBs? Artix was the commercial product, and Celtix the OSS version. Every time IONA added new function, it tended to put it in the commercial one first, and only backfitted it to the OSS version if it didn’t get wide acceptance. So, be aware that if you go with an OSS project you may have to take a commercial license in the future.

Watch out for projects that claim massive acceptance but which in truth are only supported and used by a small minority. A typical trick to watch for is claims of ‘Millions of donwloads’. This is really weak – remember that if something is free to download, every student in the world is likely to download it at least to play with. Only a timy fraction of these downloads would even move to a point of trying to actually use it.

The best tip of all is to wait for clear signs that a particular OSS project has gone mainstream. So, Mozilla Firefox is so well-known as a browser, with so many users, that it is a reasonably safe bet. LINUX has huge industry support, and rich backers such as IBM. There is no way it would ever be allowed to fall behind in the technology stakes, and because of its wide acceptance there are hundreds of companies in the LINUX ecosystem now offering support, services, training and other OSS LINUX add-ons. However, If you really want to be a trailblazer, then go ahead with unproven projects…but just go in with eyes wide open.


Posted in best practices, Imported, Open Source, OSS.