Re: CSS Device Adaptation and legacy meta tags and doctypes

On May 25, 2012, at 6:31 AM, Florian Rivoal <florianr@opera.com> wrote:

> I am not sure I want that text to be normative. By making the text normative, we'd push every browser into supporting these mechanisms, and make them more reliably usable by authors. In my mind, the goal of @viewport is to replace all these, not to legitimize them. The section explaining how to convert the meta viewport tag into @viewport is informative, and that's intentional.
> 
> On the other hand, once you have @viewport in place, supporting the
> rest is fairly easy, and it might be worth doing for the sake of
> interop.
> 
> Also, I am not sure a CSS spec is the right place to normatively define
> parsing and associated error handling of html meta element directives.
> Maybe the HTML spec, referencing this spec, could define these meta
> elements. HTML would define the parsing and what set of @viewport
> descriptors it corresponds to.
> 
> Maybe we could say that for documents unambiguously identified as
> designed for small screen, the UA must have an UA stylesheet with every
> @viewport descriptor set to auto, and leave it to the host language
> to determine what it takes to unambiguously identify a document
> as designed for small screen. HTML could (if they wanted) specify that
> the HandheldFriendly meta element does that too.
> 
> All these reservations don't really apply to the XHTML-MP doctype, though,
> so I think I'm fine with making that normative.

Surely determining which doctypes count as XHTML-MP doctypes is an HTML parsing issue, just as much as specific <meta> values. It seems like the spec defines <meta name="viewport"> in a non-normative section as well.

It seems to me that if these types of things are required for compatibility, they should be normative in some spec. That spec could be the HTML spec, normatively defining how to parse these constructs and produce equivalent CSS viewpoint rules. Or it could be the device adapatation spec itself, so all the thematically related material is in one place. Defining it in the CSS Device Adaptation spec but as non-normative seems like the worst of both worlds, since the spec mixes in HTML-specific material but on the other hand it's non-normative so there is no guarantee of interop if everyone implements the spec.

Regards,
Maciej

Received on Tuesday, 29 May 2012 00:43:20 UTC