The era of much-touted surveys that conveniently "prove" a point that just happens to match with the sponsoring company's strategy are largely over. But surveys do have a place, provided they are conducted properly and the results are not taken out of context.
For example, the new report from the Linux Foundation called "Linux Adoption Trends: A Survey of Enterprise End Users" rightly contains the following caution about its results:
Admittedly, The Linux Foundation End User Council and other enterprise end users who are motivated to complete a survey from The Linux Foundation are not an unbiased lot.
It then adds:
But the organizations' size, buying power and technical backgrounds – as filtered by The Linux Foundation and Yeoman – certainly make this data worth noting
I think that's true, even if it's telling us about the views and intentions of a very particular group – those already using GNU/Linux in their company – and not offering a statement about wider future trends for GNU/Linux for businesses in general.
The picture that emerges is very consistent: people using GNU/Linux are happy with it, and intend installing more of it, for more mission-critical tasks. Interestingly, they are even beginning to swap out Windows systems as well as the old Unix boxes. An important shift is that GNU/Linux is now sufficiently mature and familiar in this context to be seen as "more strategic" to the organisation in the CIO/management's eyes. The main driver for adoption among correspondents is Features/technical superiority, ahead of TCO and security.
Two subsidiary results struck me as particularly noteworthy. The first, is that over a third of respondents said they were using GNU/Linux on the desktop, with another 11% planning or considering it. Most of these were small deployments, but they do at least show that there is growing awareness of GNU/Linux as a solution outside the server room.
The other thing was that the top two obstacles to adoption were driver availability and interoperability with other platforms/applications. It's hard to do much about the former (and progress is being made there anyway), but the latter is an area where we can make a difference. That's why it is so important to ask politicians at national and European levels to support truly open standards that allow free software solutions licensed under the GNU GPL – in other words, RF, not FRAND, as I was discussing yesterday.
You can get the report free online, but need to register for it. That seems a mistake to me: the value of getting this kind of positive information into the hands of sceptical IT managers far outweighs the procurement of a few email addresses. If the Linux Foundation were wise, they would be truer to their community and adopt a little more openness themselves in this respect.