Then they decided that it needed to run some form of Windows.
OLPC didn't decide it needed to run some form of Windows... Microsoft decided it needed some form of Windows. To not be left out, Microsoft ported WindowsXP. All OLPC did to support that was made sure OpenFirmWare would boot Windows (and subsequently standard Linux distros) to prevent Microsoft from completely overwriting the firmware with standard BIOS code preventing Sugar from booting ever again. http://lists.laptop.org/piperm... OLPC still pushed for Linux / Sugar. All the stories I read were about the Sugar installs. Microsoft also pushed the Classmate to be a Windows platform.
Although OLPC had great intentions I feel there were several problems:
-Sugar was ridiculously slow. I know it's running on a crappy AMD Geode, but it was real slow.
-Assumption that everyone wants to be a programmer. One of the reasons Sugar is so slow is it's written in Python. Easy to modify, but being an interpreted language, it's slow. How high a priority is being able to modify the OS's GUI?
-Poor selection of apps. Poor selection of actual learning materials. Instead there's a million "learn to program" type apps, and some crappy games. Yes I think accessibility of learning to program is good, and lots of people on /. will talk about how hacking away on Basic on an Apple //, or POKE on C64 at their school got them into CompSci, you really are the minority. With the amount of money being spent on the things, they better really help with the basics of education (3 R's) first.
-Poor support of the deployments. In many cases it seemed they were dropped off, and it was up to the teachers to figure out. These are teachers not very familiar with computers, so what are they supposed to do with them?
-The platform doesn't age well with the students. Sugar is really targeted for young elementary students. I think if it was designed to have access to a standard Linux desktop (Xfce maybe? I think that's the one that hasn't gone to crap like KDE, Gnome, Unity and will run well on old junk) it would be good for older students, as well as opening up the platform to a lot more applications and resources. XO-1.5 at least was designed to dual boot Sugar and Fedora 11.
-Trying to be too much: Ground up building a new GUI, ground up building a new boot mechanism (OFW), wandering goal (XO-1, XO-1.5, XO-1.75, XO-2, XO-3, XO-4), means they're not dedicated to supporting a certain platform for a longer period of time. With the amount of money these poor countries are spending on it, it should be a solid supported platform for a while.
Really I find a lot of these problems are shared with conventional technology platforms in education. At some point TV was going to be the be all and end all to education. Nope. Growing up my school had Apple //, Mac Classic, iMac, and eventually Windows PCs. Still questionable how much they added to the educational experience. I remember playing games and typing tutor on the Apple //, but there were three of them in the back of a class of 20 students. Although I could use a word processor / Spreadsheet programs (as a commonly toted example of why computers in school are important), it wasn't till University, or later "Real world" / workplace that I learned proper way of doing things (such as styles). At the very least in developing countries any push for computers (OLPC, Tablets, etc) should be a good ebook reader first, with tons of "open textbooks" / lesson plans, but I didn't see that materialize in OLPC.
In the developed world I see it continue. Look at the amount of schools spending ridiculous amounts of money on either laptops for each child, or tablets for each child. Do they actually do anything? In my Junior/Senior year in University there were students that did their Freshman/Sophmore year at a collage that really promoted laptop use. They got assigned a laptop and used it in every class. Maple for math, etc. I don't see them being any better off than me, and I still think pencil and paper only is good for learning the basics (eg: 1st year calculus). Once you have the fundamentals, and are trying to crunch large data sets, obviously using software makes sense, and if you understand the basics, using the software is a breeze.
There is some thought, or hope, that the OLPC inspired netbooks (like the EeePC), and subsequently tablets, which are dropping in price. Hopefully some of these technologies might eventually advance education.