On Wed, Oct 3, 2012 at 11:16 AM, Boris Zbarsky <bzbarsky@mit.edu> wrote: > On 10/3/12 2:12 PM, Tab Atkins Jr. wrote: >> I don't like this. We generally force people into strings only when >> required - when the necessary name is defined by something besides the >> author, so there's no guarantee that it'll match the ident grammar >> (and making people remember the ident grammar and use escaping is >> painful). Idents are more convenient (two less characters) and more >> consistent with the rest of CSS. > > Quite frankly, I think the restriction to idents for user-defined stuff like > variables and counter style names is a huge pain in general. In programming > languages it leads to things like camelCase and whatnot that are often much > less readable than just having a multi-word string... A nice benefit of CSS over most programming languages is that it, like Lisp, can use dashes in its idents. Many people find that more readable than camel-case. > But yes, lots of tradeoffs here. If all else were equal, I'd agree that we > should use idents, not strings, because that's what people expect. But it's > not clear that all else is equal. I think you're making this to be a larger problem than it is. The easy solution (just make user-idents ASCII-ci) is trivial. ~TJReceived on Wednesday, 3 October 2012 18:22:44 UTC
This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:22 UTC