Allow us to start this entry by admitting that our heads are spinning just as much as yours. The events that unfolded over the past two weeks or so and all the responses, articles, comments, blog-posts and e-mails discussing them have been simply mind-boggling.
So some if not most of the things we're going to discuss here have probably been already mentioned elsewhere. However we do feel that an attempt to write a comprehensive text on why Windows on the XO isn't good for the educational mission at the heart of the project has to be made. And if only to help ourselves put things into perspective.
Translation and Localization
The first thing we want to discuss and a point that surprisingly hasn't received as much attention as one would expect is translation and localization. It really shouldn't be necessary to explain why translated and localized software is important when it comes to an educational project.
Now let's look at some facts: According to Microsoft "there are 24 fully localized versions of Windows XP Professional", 6 1/2 years after the operating system was introduced. Compare that to OLPC's Pootle server and you can find more than 40 languages in there. Admittedly, many of these translations haven't gotten very far, however everyone with an internet access can contribute to these translations.
Or you handle it like OLPC Nepal did and organize a
Compare that to Microsoft's approach, does anyone even know whether it is possible to translate it to languages such as Nepali? Are we going to see Translation Nights in Microsoft HQs around the world?
Next point, performance. Now this is somewhat of a tricky issue since few people have actually seen the tailored Windows XP running on the XO. Based on experience with the Geode LX800 platform running Windows XP it is however clear that it will run Windows XP just fine. As Christoph previously commented:
"Moderate multi-tasking does slow it down a bit but in general it's a very usable system for e-mails, browsing the web and office applications."Plus admittedly Sugar doesn't quite offer the fastest user-experience at the moment either. However, and we feel this is a vital aspect, over time open-source software tends to improve in both performance and stability through an iterative development process. Windows XP on the other hand tends to become slower after just a few months of usage.
Long Term Support
Moving right along to the question of long-term support. With the XO being designed for an estimated lifetime of ~5 years one might wonder how Microsoft is going to support their product a couple of years down the road. While Microsoft recently announced that "Extended Support" for Windows XP will be available until April 2014, it can be assumed that the overall level of support in terms of security and maintenance updates will gradually decrease.
The thing here is that once Microsoft decides to terminate its support for Windows XP there's very little that customers and developers can do to change that. With an open-source operating-system, on the other hand, any country could simply hire a bunch of knowledgeable developers and maintain their code-base until the end of time. Another key advantage that Sugar has over any other software solution, be it Linux or Windows based, is the tight integration of collaboration.
Now some might argue that this feature isn't or shouldn't actually be part of "Sugar", however the fact remains that re-engineering traditional systems to enable this level of collaboration would take a long time. As Walter Bender recently put it in an interview:
"...if you are going to collaborate with people, we need to make it a first-order experience."Again, none of us has seen Windows XP on the XO however it would be very surprising to see Microsoft offer anything even remotely as capable and versatile as the collaboration features in Sugar. Things aren't working perfectly just yet but we're definitely moving into the right direction.
The 'write'-activity on the XO is still by far the simplest way to collaborate on a text compared to any other solution that we're aware of. Other technical advantages that the Linux + Sugar combination can offer is the tickless kernel that aggressively reduces CPU power requirements where we don't see Microsoft catching up anytime soon.
Often Windows's power-management seems to be more effective than what even the latest Linux kernels offer, however adapting Windows XP to deal well with all the suspension / resume cycles that are happening on the XO is probably not that trivial. In fact it is our understanding that Microsoft will not modify the kernel for the XO but rather only make use of tailored drivers and software.
Another aspect to consider is that a lot of thought has gone into the overall design of the Sugar UI, especially when it comes to colour selections and the contrast between them, to ensure that the interface remains usable when relying on the XO's black and white display mode. One last point that does get mentioned a lot is cost.
We can only assume how much Microsoft would charge per license but it will probably be in the $6 to $10 range. That means that for a country deployment such as Peru the cost would suddenly increase by at least $1.5 million dollars. We believe there's many more useful things that can be done with that amount. Especially since the early reports from places such as Uruguay and Nepal indicate that Sugar works well once you actually let children use it.
Last but not least the argument of "countries would buy XOs if it came with Windows XP" is also questionable. It is more likely that many countries are waiting to see how the current deployments work out before deciding to invest their own resources into such an initiative. The real issue here isn't money but the lack of conclusive research into just how effective a tool an XO really is - but that is a discussion for another day.
In the end we hope to have given a quick overview of some of the real reasons why we believe Windows on the XO is a bad idea. In our opinion an open-source operating system on the XO offers a vast array of advantages compared to any proprietary solution. Some of these advantages might not be so visible at the moment but in the long run they're going to make a huge difference.
This comment was co-authored by Bernardo Innocenti and Christoph Derndorfer. By the way, think twice before you start calling us names such as "open source fundamentalists". Most of this post was written on Christoph's laptop which runs, guess what, Windows XP SP2...