On 12 August 2010 16:00, Nicholas Clark <nick@ccl4.org> wrote: > On Thu, Aug 12, 2010 at 12:51:50PM +0100, hv@crypt.org wrote: >> Zefram <zefram@fysh.org> wrote: > > [Thanks for listing the possible characters] > >> :For our present purpose, I'd be happy with any of >> : >> : _ ^ ~ * . >> >> FWIW, I like ^, I'd be happy with ~ or *, I dislike . and _. >> >> I quite like the mnemonic sense of ^ as "starting with a clean sheet", or >> "anchoring [the flags] to a known set", though I wouldn't suggest trying >> to mention or explain that in the documentation. > > Or "from the top", and the top of the page is above instructions about > change-this-from-the-default. > >> I find all of ^ ~ * to be sufficiently visually distinctive (at least in >> any font I'm used to using). I don't find any mnemonic argument for or >> against ~. I find the quantifier sense of * slightly distracting, in that >> it seems to contradict any mnemonic sense for its use in this context. >> >> I dislike . because I don't find it sufficiently visually distinctive, >> and because it's "any" sense seems to contradict any mnemonic sense for >> its use in this context. >> >> I dislike _ because it is a word character, and because it is not >> distinctive enough from -. > > I realise that this is a bit of a bikeshed issue, but I agree with Hugo, > pretty much for the reasons that he gives. Me too with the exception that I am vehemently against using * as it is used for "backtracking verbs" and also for quantifiers. I used it for backtracking verbs because it shouts out "HEY IM DOING SOMETHING FUNKY HERE". Also I think (?* is unwise, as *? has a special meaning that is totally different. > I don't find ^ *that* confusing > > 1: it already has a second meaning, inside character classes > 2: I already look with suspicion at any ^ used at any position other than the > start of a regexp, as things like /(^beer|pie)/ don't do what some people > expected Agreed. These days I tend to avoid ^ even when it is correct and use \A instead, except when I really really mean to use ^. Especially these days with loads of people brainwashed by bogus PBP arguments that all regexes should have msx enabled. Ugh. cheers, Yves -- perl -Mre=debug -e "/just|another|perl|hacker/"Thread Previous | Thread Next