Quote:
Originally Posted by Turtle91
Any compliant renderer is required to gracefully degrade on an unknown or incorrect css. If it ignores the entire css then that is a bug and should be reported as such.
|
Nope, CSS spec allows any syntax error in a rule to invalidate the entire rule. An unrecognized property could certainly be considered a syntax error, although most renderers just ignore it.
Quote:
|
A non-css3 renderer should likewise skip over the css3 ...
|
A property with additional values in CSS3 than in CSS2 can also be considered a syntax error if a CSS2 renderer sees one of those new values, and this happens in quite a few renderers.
Even if a renderer does just ignore the unknown properties/values and treats them as no-ops, you still get things that a nowhere close to "gracefully degrading". Take your example code and remove the "overflow: hidden;" line to simulate a renderer ignoring it. Not very pretty, is it?
My primary reading software uses a terrible HTML/CSS implementation (but I'm stuck with it until there's a 10" reader that worth upgrading to), so I'm very aware of such limitations, but if (as Jon points out), baseline RMSDK (which is miles better than what I use) doesn't support something, then you really have to assume that a lot of reader software isn't going to, either.