A long time ago I was considering subversion to replace CVS. At the time in checkouts it was storing uncompressed copies of the source files. So every checkout was (approximately) twice the size of a CVS checkout. That gives quick "svn diff" and things which is nice. But the whole git repository was (as it happens) smaller, and that gave lots more fast operations ("git log", diffs for arbitrary revisions, etc.). No contest really.

A quick experiment suggests that that might still be true. The format seems to have changed (only one .svn now, and the file copies seem to be in a database or something) but the size looks similar. So for libcxx (from llvm.org) I can get an svn trunk checkout for 70M, or I can get a git checkout (containing all revision of all branches) in 86M.

subversion looks like a win if you have large binary assets. That's a case that git doesn't handle well. For typical software development it still doesn't look competitive to me.

source navigator still works OK. doxygen can provide reasonable callgraphs (especially for C). DXR has improved dramatically recently---it's not just for Mozilla now. DXR requires the code to be compilable by clang, but doing whatever's necessary for that might be a useful exercise.

Here's my idea: run two sites, the first one's like this (previously unknown to me) quirky, a good way to get candidate ideas some of which would be profitable (but as noted, a horribly ineffective way of actually doing anything with the contributed ideas), and another much more like Bennett's which can effectively filter and exploit the ideas. Then feed the latter from the former, and (with any luck) get profitable products which can then be sold. Or maybe just sell the ideas/patents (along with the list of interested customers) to third party manufacturers (or some combination).

Obviously it would be bad if people at the quirky-like site knew about the other one, so I'd want to keep it quiet (and probably vice versa). Maybe have it only operating in a foreign language in a foreign country (perhaps more than one).

(I'm not suggesting this is how quirky (or any other site) works. Seems likely they started with what seemed like a good idea and for some reason haven't been able to make it actually good (Bennett's analysis seems sound to me, what they're doing seems ineffective). And they don't care yet because they're getting press (apparently, anyway) and perhaps VC. Or maybe it's basically a hobby, or makes enough from selling the rejected ideas back to the proposers, or something else I just haven't thought of.)

Presumably the price is just the predictable "let's see who wants it before Christmas" price, though. They surely don't expect many sales at that price.

See BMJ 2008;337:a2533

Quite. So to buy/use one of these ebooks in this "open" format I need Adobe Digital Editions, available only on Windows or Mac. There are surely worse formats, but this one doesn't seem especially free software friendly.

The story suggests the article appeared "in all the news groups on the AT&T/SBC News Server". It also appeared throughout the world,
due to a typo: "Distributrion: internal".

Back when I did a maths degree in the UK, and probably now, A-level syllabuses varied wildly in what they covered. And students may (as I did) have taken two maths A-levels rather than just the one. So some students might well simply not have encountered any basic statistics, or might have inadequate knowledge of geometry, etc.

Hence the remedial tests and classes. It's not an examination, as such; it's just to see which students might benefit from having bits of the basic knowledge filled in, so only they need to sit through what would be exceedingly boring classes for people who've already covered the basic stuff.