Lee Daniel Crocker wrote:
That seems to be a growing consensus. The remaining questions, then, are the exact details. Exactly which characters do we want to consider "punctuation", and under what circumstances? My suggestion is this: After parsing URLs the way it does now, if the URL ends with one, and exactly one, of the characters period, comma, question mark, or exclamation; then remove that character and assume it punctuates the sentence. Otherwise, leave the URL alone.
There's also the endings ".)", "?)", "!)", and "...". Getting too complicated? I agree.
So the rule should be a string of *any* length at the end of a URL consisting of characters that (should shouldn't be hard to determine) aren't commonly used to end a URL. Annoying if you want to site http://www.com/page.html?secretcode=*%.)?!@? Yes, but that's going to be an extremely rare occurrence, and anybody that does that will know that they have a weird little URL and will check the <Preview> if they're not a fool. In contrast, everybody and their mother is going to write "(For more information, see http://www.com/moreinfo.html.)", at least while they're newbies.
-- Toby Bartels toby+wikipedia-l@math.ucr.edu