There are going to be wiggles when ever you have lines of code replacing other lines of code. They thought they had this covered when they put a 'pretend area' (hot test bed) up simultaneously. Not so much.

As I understand it (and know I could be way wrong) the offending snippet of code existed in the HF19 version that ran more or less happily for year and a half. It was when enough witnesses started running HF20 that the problem arose.

That isn't much of a problem, far as I am concerned. Shit happens. The fact that it took the witnesses all day to agree on and go to a fall back position is a bit scary to me.

There is a problem with everything that is touched by computers. The stuff that was written 40 years ago.

You see, nobody writes brand new code. They always put their stuff on top of another piece of code that somebody wrote in Fortran or GW Basic. So by the time we are where we are today any bit of code that makes stuff work on the innerwebs is tens of thousands of lines of code. Nobody can check them all. Ever.

It's why 'open source' is so important. When there is a glitch in open source software a thousand geeks drill backwards on it until they get to the Fortran and then replace the Fortran so you have an entire block of new code.

With proprietary software you have the company team throwing a patch on the problem and hurrying on to the next one.

Ask Microsoft how that has worked out for them. They promised Win8 would be entirely new code. Win9 they skipped and got Win10 out with a lot of new code. Late and expensive. And still related directly to the OS Gates and Allen devised in the garage.

So for me, it's not about a glitch happening, it's about the amount of time it took to get to the fallback position. I'm guessing it's time to adjust some of my witness voting.

Congratulations to your winners. You know I read every word before I answer so I knew I was on a tangent when I answered. You also know I don't do many contests so the only part of the rules I followed was the word limit. :)