Originally Posted By: Gordon Scott
Some semi-random thoughts on the update process.

I believe PG have a number of beta testers and I suspect that they're all very experienced BIAB users, who know there way around the program and know how things should work. That's a two-edged sword, because they also know what _not_ to do so that they can keep themselves out of trouble. It's very likely that, probably subconsciously, they're avoiding bugs into which we less familiar folk can crash. In a past business we had a guy who was, almost pathologically, deliberately naive when testing software. He would do the craziest "why would anyone do that?" things, but he found loads of bugs.

Quite a lot of complex software packages have multiple risk-level option, e.g., Development, Beta, Release Candidate' Release and Long-Term-Support. That is more work to maintain, but not as much more as one might imagine. Whenever I finish a bug-fix in a mature version, typically I compare the old with the newer and propagate that same fix when I can. Unusually for me (I'm a Unix/Linux man), I use WinWerge for that as I find it extremely friendly to use.

As a user, I generally avoid getting the latest new release, as there are almost always lots of bugs. I have no desire to inflict the pain upon myself; let others do it. Either they don't mind the pain, or they'll learn in time.

In my few updates with BIAB, they've usually seemed to me to be pretty shockingly buggy. But still I was tempted in by the upgrade offers. I'm of the opinion that if your new release is likely to have many bugs, then encouraging large early uptake seems like a good way to bring a tidal wave of bug reports, anger and frustration upon oneself.

One of the things over the years in my own electronics and software work, that has both frustrated me and satisfied me, is the comment "Oh, that's really easy, even I could have done that", when one has spent many hours, days, weeks, thinking very hard about _how_ to make it so easy. I try to make things consistent everywhere, even in places where I think the user will likely never go, because if they do, the product will likely behave as they expect. I personally am of the opinion that, longer term, it also actually saves _me_ time, frustration and effort.

The best systems tend to feel like there was one system architect and one vision. It doesn't matter if it's really quite a large team, provided they think together and act together. BIAB feels like one or a few people are given things to deal with and they do so in near isolation. I suspect that is actually _not_ the case, but it feels like it.

Gordon, these are brilliant insights! I hope PGM reads this and takes it to heart!