Wednesday 25 April 2007

Desktop Linux - distribution fragmentation

I've had an intention in mind for some time now to seriously investigating desktop Linux for professional life. This intention has been driven by a number of factors, but until the advent of Vista gave me a vision of the future, the intention remained just that. Why?



From a personal perspective, it could be because I'm too old or too dim (or both) to go through a great deal of education and re-education trying to work out which distro will provide adequate return on the investment in time and effort. It seems to me that the first port of call is the question "Which distro?". With DistroWatch. recording 350+ distributions it is far too easy to be intimidated out of the market. It gets a little easier when one starts to realise that actually they are not all independent of each other - Ubuntu, Linspire, Knoppix are, for example all based on (extensions of) Debian. Aha! that helps a little. The picture becomes even clearer with a graphical representation of the family tree and a text-based tree here.



So now I have a clearer idea where the distro's are coming from but it doesn't asnwer the question "Which one?". Next port of call was the Linux distribution chooser. Depending on the variations of answers to the perhaps over-simplistic questions I got pointed towards OpenSuSE, Fedora or Ubuntu. But WHOOOOOOA! while trying to get my head around the the distro thing, I've ended up with RSS feeds from a dozen or so sources. It doesn't take much to realise that Novell and OpenSuSE have incurred the displeasure of at least a section of the 'community', and Red Hat has its fair share of critics. The last thing I need is to spend a lot of time assimilating the vagaries of a distro that may be marginalised by the wider Linux community. Ubuntu it is.



An Ubuntu it has stayed so far.  It does (almost) everything I wanted it to, and certainly everything I needed it to (not always ideally, but I'll make do). I have a server running on an old laptop, a desktop and notebook running as workstations; Ubuntu itself is going from strength to strength; Canonical has the philosophy right in term of community (although at some point in the future they are going to have to find a way to turn a profit), and I see no compelling reason to move away from where I am. I can get on with 'real' work with a feeling of comfort that the tools I use will be OK for the next while.



That's great. I'm going to recommend Linux to my old Dad. Not.







Powered by ScribeFire.

Saturday 21 April 2007

I am not an enthusiast

I am not an open source enthusiast. That is not to say that I am not enthusiastic about open source, bacause I am, but I am not an enthusiast in the same way that I am not a car enthusiast. I do not spend my spare time tinkering with the hood up adjusting petrol mix for optimum performance, nor do I subscribe to glossy magazines informing me which form of alloy trim is in vogue. Don't get me wrong, I don't have anything against those who do, it's just that I don't.

I have been an IT professional for over 25 years. I have been in my time developer, sysadmin (many OS's), db admin, hardware support, cable puller and all the rest. I have no issues with doing any of that stuff, it's just that it's not what I do now. When I'm doing what I do now, I'd like computers to just work. And for that matter, when I'm not working, and I want to play some music, I don't want to have to google the latest encoding schemes and determine which codec I need to update. I want it to just work. Therein lies a hurdle which Open Source has negotiated poorly to date - one of product management and marketing. And the reason that that hurdle remains is largely in the makeup of the FOSS community.

Open source communities are made up of enthusiasts as a rule (yes, there are exceptions to the rule but not many) and it is inevitably (and to some extent, rightly) that community which determines what happens to a product - which features, bugs and developments are prioritised. Enthusiasts in the community, though, not only don't mind tinkering under the hood, but often prefer a product which requires tinkering. The result is a package which looks great from a developer's and enthusiast's point of view, but not necessarily from an end user's.

Why does it matter? It doesn't as long as FOSS remains bounded by its communities. But there are increasingly commercial enterprises working on a variety of business models. Ubuntu is a good example. Cononical clearly would like Ubuntu to be considered the desktop Linux of choice and would like to break clear of the community boundaries. In my opinion, though, the packages that comprise Ubuntu (and other distributions) are not yet adequately focussed on what that target audience needs to make the switch from Windows.

Of course, many aspects of Linux based desktops (including Ubuntu) knock the socks off Vista. But many are little better that Windows 95, and for that target audience it isn't enough that a package will just about do the job, eventually.

It isn't for nothing that organisations like Microsoft spend millions on market research, focus groups, usability testing and the like. Early on in my entrepreneurial career I was told a truism which I have held central ever since, that the best way to ensure you sell what you make is to ensure you make what you can sell. If FOSS wants to break into the non-enthusiast market, then this must be addressed. Even the otherwise excellent book 'Producing Open Source Software' by Karl Fogel (available here) barely mentions the marketing function.

But how could marketing functions be included in the OSS process? Should the marketing itself be considered to be as open as the source? There may be a platform for opening up the information in a Creative Commons license, but the framework under which the information is derived is unclear (to me anyway). I think it likely that that framework will be different to traditional commercial products.