I attempted to get Sugar running on a computer lab I was teaching in at the time, and it was a fantastic idea, but had enough bugs for it to be effectively unusable.

> would prepare 3rd world kids for exactly no other computer

The point wasn't to teach them how to use computers, otherwise they would have just shipped a Windows device. The point was to make an educational tool; even using it as an ebook reader is a big thing in places where there are no textbooks.

They could have just used or written something that introduced them to standard UI metaphors instead of giving them Sugar. Probably would have saved effort.

Is there anything that great about standard UI metaphors? Most of them are descended from designs when hardware and performance was severely constrained, and hand around because it's too hard to change everything for most people (touch devices like smartphones and tables are the main exception, but they were only just getting a foothold at that time). Given the opportunity (in this case, since the audience had little or no prior expectations of a UI), why not try something new?

Because it works wells enough for education purposes and would help prepare them for computers that many of them would encounter eventually. Sugar was well intentioned wankery. You could have had special use case software for off the grid lessons, but giving them all a bizarro UI that wasn’t like anything anyone had ever seen before didn’t make a whole lot of sense. I’d say it was one of the big reasons OLPC failed. Also, had they forked off of existing DE/WM around at the time it would have been less buggy and would have involved less effort on their part. Sugar dropped me to console more than once, and I don’t expect most students or their teachers knew what to do besides reboot and probably lose work at that point.

Interestingly, in Paraguay the kids who did use their XOs much at all generally installed another windowing system. There was a nice easy package that even not-very-tech-savvy kids could install, I think put up by a hacker group in Uruguay (I'd have to check my notes on that one) with instructions in Spanish, that also came with some simple videogames and a video/audio player. Sugar was really not designed for media consumption, and that's what interested many of the kids!

> IIRC there were no moving parts (fan? don’t recall), which was another huge plus.

Didn't it come with a hand-crank for powering the laptop? Surely that should count as a moving part.

It did not. The OLPC project wanted this feature, but it proved impractical to implement.

And what was the excuse for not including a solar panel on the back of the display to let the computer recharge in the sun?

The OLPC was a small device -- the usable space on the back was roughly 5" x 8". Even nowadays, this would be a very small solar panel; with early-2000s technology, it'd be completely useless. The project experimented with larger external solar panels costing $50 or more (roughly the same cost as the entire laptop!) and found that they were minimally effective for charging: https://wiki.laptop.org/go/XO_Solar

Because when the screen is visible to a user, the back of the screen is pointing at the ground, not the sun.

Here's a video of an external panel running the laptop: https://youtu.be/ITHNbOrPQyM

The idea is, when unused, the laptop could be left in the sun, where the back of the screen could trickle charge it without any user intervention.

Ideal? no. Better than having no power or using a crank? yes.