Downsides:
1) If you care, loss of credits if beta WU's error out. (Though the same occurs when production WUs error out too.)
2) Frustration, as you can report a problem and the faulty code gets put into release anyhow.

It's not any work for the user. It would be nice if it was (like, if they did some follow-up and wrote some test programs to work out where the problems were) ... but sadly not.

No, it's not hard to track down. At least, it wouldn't be hard, if they'd actually bother to communicate (do they know how to do that? -- they're programmers ) with the people running the WUs. If they isolated the faulty code and made test programs available so that the people whose machines crash could run them, such problems might be relatively easy to track down.

The more you guys (sorry you're in the middle, honestly) defend the incompetence, the more I'll shoot it down.

Yes, WCG isn't new to this, which is why it's amazing how poorly they're handling it. Yes, it is easy to criticise, because they put a program into beta test which crashed, then put it into production still unfixed, and still haven't fixed it in 3 months. Yes, it appears criticism doesn't help, because they're not taking any notice of it. I'm more than happy to let them do their job ... whenever they start doing so. If you have any solid evidence that they have made any real attempt to fix the bugs they don't appear to be bothering to fix, by all means say so. Otherwise, your suggestion has less substance than my criticism.

The original question here was about what beta testing is like. I feel it's more than appropriate to answer the question. If you wanted to contribute to the answer of the question in this thread directly, rather than criticise me, you might simply say that not only will the techs sometimes fail to fix problems, but that the "helpers" on the boards will then defend their inaction as well. If WCG were a company with that attitude, I would definitely take my business elsewhere, as it's completely unprofessional.

The project programs themselves were not written by WCG. WCG is only supplying a shell or might have to convert the program for personal computer use. If there is a program bug, as opposed to a BOINC bug, it would have to go back to the project submitter to repair. It is also possible that there are more different types of platforms running here than the submitter can even replicate. Maybe the submitter doesn't want to fix something that only happens less than 1% of the time.

(That is only idle speculation.)