- From: Simon Pieters <simonp@opera.com>
- Date: Thu, 23 Apr 2015 10:18:16 +0200
- To: "Tab Atkins Jr." <jackalmage@gmail.com>, "Florian Rivoal" <florian@rivoal.net>, "Chris Lilley" <chris@w3.org>
- Cc: "www-style list" <www-style@w3.org>
On Wed, 22 Apr 2015 23:04:03 +0200, Chris Lilley <chris@w3.org> wrote: > Hello Tab, > > Wednesday, April 22, 2015, 9:21:31 PM, you wrote: > >> 1. reset (already suggested by you, just putting it here for >> organization) >> 2. ua-default (suggested by zcorpan) >> 3. user-agent > >> I like ua-default. It's even clearer than "default", and makes it >> really obvious to people what it does. It also seems to be >> practically guaranteed to not already be used by authors in any >> <custom-ident>s we have. Its only downside is that it's not >> technically correct - when used in an author-level sheet, it causes >> the user stylesheet to be applied too - but I don't think that's a >> complication that matters. > > For that reason I prefer reset. Because the prose can explain what > exactly you reset to. > > Instead of spec text that says "despite what the name might lead you > to believe, this value also ..." For this to make sense, you have to assume that CSS authors typically read the spec before using a feature. In practice, that almost never happens. It is much more useful to have something that most CSS authors understand well enough to be able to use it as intended without having to read the spec or other documentation. If we have "reset" and "unset", I imagine people will confuse the two and will have to look up every time which one to use. I consider that to be a language design failure. It's like replaceChild(nodeA, nodeB) in DOM -- it is ambiguous and impossible to remember what the argument order should be, so I have to look it up or trial-and-error every time. -- Simon Pieters Opera Software
Received on Thursday, 23 April 2015 08:18:57 UTC