I think the gatekeepers of the institution feel that human knowledge is not increased when homework answers are published on the web, rendering the assignments obsolete for future instances of the course.
It was easy to cheat 8 years ago (from what I remember), it's surely easier to cheat now, and it's certainly not because this one kid stood up for something in which he believes.
(Edit: I don't have all the answers, or an opinion on this piece, just playing devil's advocate)
Although, honestly, people are going to hand down homework answers no matter what. Professors simply should assume that if the question isn't new, an answer has been made available to the students somewhere at some point.
Nowadays it is hardly any extra work for the prof either, any coursework of substance gets put through plagiarism detection anyway.
Also, after "Intro to X Programming Language", there aren't going to be any trivial assignments and every (or nearly every) solution should be different. At NCSU they had software that would examine all code for assignments to detect people sharing code and trying to make trivial modifications such as changing variable and function names, and changing the order of code. I had quite a few friends get caught by it.
The point of my rant is that some students are going to cheat no matter what. Having the solution posted online isn't going to really change how many people are cheating if it is made clear that the professor will be looking for cheaters.
tl;dr: "don't cheat, we'll find you" -> no effect on cheating. "These are the tools we'll use. Here's how they work" -> massive decrease in cheating.
1. write a sed script that strips out html-tags from arbitrary, well-formed HTML.
It's pretty hard to fail to notice that 5 people used the one-liner from http://www-h.eng.cam.ac.uk/help/tpl/unix/sed.html, but can't explain what the hold-space is. Also hard to hide the fact that you used the hold space!
Believe me, there are many better things he could be doing with his time than re-creating exercises....
Some people are currently exploiting the "efficiency" you mention, which is in my opinion a bigger problem than the extra time needed to create fresh materials.
What course could you imagine where it would take more than a few hours to create a fresh exam or batch of quizes? We're talking about undergraduate classes and any prof qualified to teach them could easily rattle off dozens of great (and unique) questions on a moment's notice.
So not only do I not think it would be a waste of time, but it would result in fairer tests/quizes, less exploitation, and fresher and more topical material.
Why do students use archived answers in the first place? Apparently, to get credits for (and ultimately pass) a course whose course points they need for something but whose contents they don't want to learn that badly that investing their own time would seem worthwhile.
I see two ways in which to attack this situation:
1. Remove the possibility of passing a course without having understood its contents: e.g. oral exams with "why" questions.
2. Remove the perceived necessity to get a good grade: on the small scale, the easier it is to cheat in a single test, the less it should contribute to the grade. On the larger scale, perhaps curriculae need to be more diverse such that a "I just want to be a Java programmer with a degree such that I earn more money" student can become "just a Java programmer with a degree in just Java".
If the professor has to constantly come up with brand new problems, the students are forever seeing rough drafts of unvetted exams.
Here's a very simple example of a failure I had that illustrates the kinds of things that can unexpectedly go wrong: "what is the angular speed of the second hand of a clock?"
After struggling for a long while with this question, a very bright Asian student finally asked, me: "I see three hands on the clock: obviously they are the first hand, the second hand, and the third hand. Can you please tell me which one you are calling the second hand?"
At the graduate level it may be easier (I've never written a graduate exam) because, fuck 'em, they're grad students, let them deal with it.
In my experience most such questions are simply intended to verify a cursory familiarity with the subject -- or to be easily graded. Easy grading is another aspect of laziness that ought to be done away with.
Why must an exam be multiple choice rather than essay answers? Only so it can be chugged through a machine so nobody has to think while grading it.
I have a friend who was an art student. Every major assignment was put into a public portfolio for future employers (etc.) to look at. This is considered standard for the field. It's also possible that someone else could look at her portfolio, copy her assignment, and turn it in as their own. Despite that, it's still her right to share her creative work with the world.
As programming is just as creative as design, I don't see why programmers should be held to some other standard. We need jobs out of college too, and we are also proud of our work and want to show the world.
If this makes the professor's life harder... well, that's so sad that I'm going to cry.