Hell yeah, we love doing it. Keeps us fresh and interested in the ongoing development of the project and helps push other folks forward with new web patterns, tech, etc.

It not only gets rid of clunky code holding you back, but often the code written to keep legacy working is a huge attack vector because it's generally not scrutinized as much. because well, how many people using it really care if the little chunk of code to make it work in IE6 isn't super solid? Not many... how many 0day hunters care? Many of them.

Dropping old tech means projects/companies that require that support have to do it in-house. We need to support IE7+ for much of our UI, unfortunately, because of paying customers in hospital environments where upgrading is a nightmare.

So it's a good thing in that libraries can be leaner/meaner for everyone who doesn't need to care about old tech, but it also means the unpleasant work of supporting the old tech now has to be carried on by lots of different developers in lots of different companies, rather than being done well in one place.

If there's any possibility of keeping alive an "old-browser-support" branch of BS (like jQuery's 1.x and 2.x streams), that's rather better than abandoning it -- in theory we might be able to contribute some developer time into it.