> Let hook something up to the phrase query, but I would just suggest another > character so that it's not confusing that the same operator is used for two > different concepts.
I like using ~ because the concept of "slop" and the concept of "fuzzy" are pretty similar in most users minds, I'd think. Also, there's some value in keeping the set of characters we "usurp" from the space of valid query text to a minimum. And the ~ symbol is commonly use to mean "approximately", so its semantically sensible as well. I do NOT want this to degenerate into something perl-like, where $| means one thing and $/ something else, with nearly every punctuation character taken. > Really I would just suggest excluding what is currently used and ? (used in > urls), & (used in urls), > (would have encode for xml), < (would have to > encode for xml), % (can use be used to escape characters). You make a good argument for overloading here -- a lot of symbols have semantics in some domain or other, and we don't want to interfere with those. -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
