A few years ago I was invited to join the Esri beta test for 9.2. I declined because "I already log an issue for every x hours using ArcGIS why would I want to do even more? I need to do my real job too.". I reconsidered for 9.4 (renamed 10 for release) and joined the beta. It was interesting and cool to see the new features, but in the end I didn't participate much for exactly the same reason I declined earlier, it got too much in the way of doing my work.

On average I log one case with tech support for about every 8 hours of arcgis work (I don't use it full time), and for every case I document and log there are two I don't bother with (too hard to explain, too hard to reproduce).

To be fair, probably a third of the cases I log can be attributed to user error (I did something wrong), something specific to our environment (they can't reproduce), or it wasn't arcgis at all (it's an addon/windows/msoffice/foobar bug, we can't help it).

Lest this be perceved as a wholly "arcgis sux" rant, I encounter and log issues with most of the other software I use also (excepting Microsoft products, as far as I know there is no public channel for that). The incidence of Arcgis cases is so high because, a) I do use it a lot, b) we pay mucho denari for tech support so I'm darn well going to use it, c) Arcgis is big and complicated, so there are many many opportunities for bugs to live.

On the other hand to my recollection I only ever logged feature requests for Arcinfo Workstation and no bugs. Of course I didn't know as much then about how to identify and document the critters. Or maybe I was just nicer. ;-)