[_delirium]

That seems reasonable; I didn't actually mean it as a criticism, just an explanation. Publishing only in French has downsides for readability outside France, but as you say, publishing only in English has downsides for readability within France. Most CS researchers and students I've met would prefer the English solution, mostly because they feel that, unlike in fields like philosophy or political science, the average French person isn't interested in reading their work anyway--- the only person who's going to read a theoretical CS thesis is another theoretical CS researcher. And in that case, they'd prefer the international community of theoretical CS researchers to be able to read their work.

Denmark is an interesting example of the opposite case: there are now graduate degree programs where you cannot study in Danish, because all courses and coursework are English-only. That's controversial to some extent, for obvious reasons. It does have some upsides from an international perspective, since Denmark can now hire researchers who don't speak Danish (which is why I'm in Denmark currently), and can also accept PhD students from other countries without requiring them to learn Danish first. But the situation differs from French because Danish has many fewer fluent speakers (about 6 million), so works published in Danish reach an audience much smaller than works published in French do.