> I do not understand how anyone could tell if the kids are learning anything if they are not tested.

So, we have no evidence that you know anything about programming language design and implementation, because you've never taken a test on it?

There are various ways of demonstrating mastery. Tests are one -- and I never suggested that there shouldn't be any; only that they shouldn't be the sole factor guiding the curriculum. Papers and other kinds of projects are also very important. They're somewhat out of favor, because they're more difficult to grade, and especially more difficult to grade uniformly on a large scale.

But I think the most common way that people demonstrate mastery is by simply getting interested in material at the next level of difficulty. Kids, if they read at all, will naturally read at a level that challenges them at least a little, because anything less is boring. Look at what level the kid gravitates to, and you have a pretty good idea where they're at. (Some will push themselves harder than others, of course.)

> So, we have no evidence that you know anything about programming language design and implementation, because you've never taken a test on it?

I wouldn't be offended if I applied for a compiler job and I was asked questions about compilers. In fact I was, when I applied for a job at Microsoft a long time ago.

I don't understand how someone can master a subject and yet be unable to answer questions about it.