Footnote output not valid.

[Waylan Limberg]

FYI, I found an examples page in the microformats wiki:
<http://microformats.org/wiki/footnotes-examples>

Because of the inclusion of the 's' in the url, I didn't find it right
away. In any event, it documents various existing examples of footnote
implementations (including Gruber's). Interestingly none of them use
"rel" attributes (or "rev" for that matter). Maybe we should just drop
that usage altogether in markdown.

On Wed, Aug 8, 2012 at 3:35 PM, Waylan Limberg <waylan at gmail.com> wrote:

> On Tue, Aug 7, 2012 at 10:30 PM, Michel Fortin <michel.fortin at michelf.ca> wrote:

>> Le 2012-08-07 à 15:43, Waylan Limberg <waylan at gmail.com> a écrit :

>>

>>> Of course, following markdown.pl's lead, the default is XHTML, so

>>> yeah, most will never get HTML5 output. Although, I suppose we could

>>> easily have our footnote extension not add the `rel` and `rev`

>>> attributes to footnotes when HTML5 is selected.

>>

>> Well, I'm not so sure rel=footnote is actually almost valid. The HTML specs gives a few normalized values, then says that other values must be registered on this microformat.org page:

>> <http://microformats.org/wiki/existing-rel-values>

>>

>> rel=footnote is there on that page, in the POSH usage section, citing Markdown preprocessors as the source. But it's missing a spec. Someone would have to go through the microformat process and create a spec for it and it'll become valid.

>

> True, but it is also listed in the "dropped" section. If one it

> proposing a spec for anything that has previously been dropped, there

> is a larger hurdle to jump. For example, the proposal needs to explore

> why it was previously dropped and why that reason does not apply to

> the current proposal. And if I understand it correctly, certain

> reasons (for previously being dropped) will cause all future proposals

> to be rejected automatically, while other reasons may be open to

> reconsideration. Unfortunately, whoever added "footnote" to the

> "dropped" list left the reason as "unknown". So it looks like it might

> be a lot of work to submit an acceptable proposal.

>

>> Rev is definitely deprecated however. I don't think it'll come back. If someone defines a spec for rel=footnote, it could also define what is the opposite of rel=footnote and replace the rev=footnote with something else such as rel=footnote-ref. Then the path forward might be to include both rev=footnote rel=footnote-ref, unless someone has specified a flag to prevent the rev attribute from showing up.

>

> But until that happens, our markdown implementations are outputting

> invalid html. And according to the page linked above, they should not

> be used at all meaning we should be removing any occurrences of

> "rel=footnote" from our markdown implementations until such a proposal

> is approved. Of course, that only applies when outputting HTML5.

>

> --

> ----

> \X/ /-\ `/ |_ /-\ |\|

> Waylan Limberg




-- 
----
\X/ /-\ `/ |_ /-\ |\|
Waylan Limberg