Propdef grammars should link each component value to its definition

[fantasai]

Compare this version, in which each keyword or function in the propdef links to its property-specific definition
http://dev.w3.org/csswg/css-fonts/#font-variant-prop
to this one, in which only <valtypes> get linked to their definition
http://dev.w3.org/csswg/css-fonts/Overview.html#font-variant-prop

[fantasai]

(This might require additional markup support from the author.)

[tabatkins]

There's nothing to do here on the Bikeshed side; it already has all the capabilities necessary. I just marked it up differently than the original spec did.

[fantasai]

No, this is not fixed. E.g. the 'ordinal' keyword in the propdef does not link to its definition, as it does in the non-Bikeshedded version.

[tabatkins]

Again:

There's nothing to do here on the Bikeshed side; it already has all the capabilities necessary. I just marked it up differently than the original spec did.

[fantasai]

Can you mark it up in the way the Fonts editor wants it to be marked up then?

[tabatkins]

We don't do that for any other spec.

[fantasai]

I'm not sure what's the problem here. jdaggett's linking scheme (each keyword in the propdef to its definition) is very clearly more useful. This is especially true for the Fonts spec, but I don't see why it wouldn't be more useful anywhere else. It seems to me that at least in most cases, we have enough information to do this linking. No?

[tabatkins]

Possibly. Sometimes the important grammar term is a longer expression than a single keyword, but yeah, single keywords could probably link up automatically.

[fantasai]

Right. I think what's needed here for parity is single keywords linking up as well as function names linking up. Is that fair?

[tabatkins]

I should be able to do that. Open up a new issue for it?

[fantasai]

Um. That's exactly what this issue is about. Why not just reopen it?

[tabatkins]

Indeed it is.