begin quoting Lan Barnes as of Thu, Aug 24, 2006 at 06:32:54PM -0700: > On Thu, Aug 24, 2006 at 06:11:36PM -0700, Stewart Stremler wrote: > > > > He'd also want a quick and easy way to abort the sending. How would > > he do that from Thunderbird? > > As I read this thread, I'm left with the unixer's question about > deletion confirms ... how many do you need?
I want to say that if you need to confirm a deletion, you've messed up your design somehow -- you shouldn't query for deletes, you should deny them unless you've pointed out that you really do know what you are doing. Thus, :q in ex/vi doesn't query "are you sure you want to quit", but says "yo, dimwit, you changed something, if you want to quit, use !". And rmdir won't delete a non-empty directory, and rm won't delete a directory at all without the extra option. > I've been stung by immediacy like everyone else, mostly in deletes. But > is there an alternative that isn't embarrassing? God, I hate saying > "yes, yes, yes" to windoze ... and it doesn't make me any more careful, > alas. I know. But the query, query, query-again algorithm is so *easy*. And it has such wonderful metrics in the cognitive design space. > I'm not saying it isn't an occasional bump. I just can't figure out an > algorithm that can tell my MUA which mail to hold back. Hold none, or hold 'em all until approved. But really, the problem is mostly with mailing lists. Newsreaders nailed down many of these issues years ago. We're sticking with the "good enough" solution, so we can expect to have a few rough edges to put up with. -- _ |\_ \| -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
