Specification: http://www.w3.org/TR/html5/
Section: http://www.whatwg.org/specs/web-apps/current-work/#top

Comment:
Hello,

I'm just writing a quick note to question the non-inclusion of the "feed" link
type (rel="feed") which I remember coming across a few months ago.

When I originally saw the new rel type, I was very glad to see that there was
finally a way to specify a feed for automatic browser detection that didn't
declare it (erroneously) as an "alternate representation" of the document,
which had always seemed to me somewhat of a hack caused by feed technology
overtaking the existing markup standards.

In that respect, the new link type seemed like a good, semantically specific,
and progressive addition to the language.  It is already supported by Firefox
at least, and since discovering it, I have been using it on all of my (X)HTML5
sites.

I see that the definition of the "alternate" link type has been specifically
updated (presumably to address the non-inclusion of "feed") to say "not
necessarily syndicating exactly the same content as the current page", but
this seems like a very uncomfortable way to cram multiple, potentially
ambiguous uses into one keyword (remember, some feeds - especially on blogs -
really are alternate representations of the current document).

From a quick Googling of the issue, the best information I can find suggests
that the "feed" type was not included due to poor browser support and user
adoption, but I would beg to differ on both points.  Firstly, as mentioned
before, Firefox (which is, after all, one of the most widely used browsers)
already clearly supports it; and as for adoption by users, I can only put
forward that I, as a developer very supportive of good standards and
semantics, personally started using it as soon as I found out about its
existence, but as there is very little information about it available yet, it
took me some time to come across its existence at all, and I suspect that the
reason for its lack of widespread adoption is simply that it is, after all, a
new feature in an as-yet unfinished standard.

Moreover, it seems to me that there is no real reason not to include a
semantically useful new rel type, especially since as far as I am aware, there
is no physical limit to the amount of "allowed values" that can be included in
the spec.  Indeed, there are already rel types included that seem to be of
very little real semantic value ("nofollow" springs to mind as a value that is
really only of use as a somewhat crude method of discouraging search engine
spam, rather than providing any useful content for a user).

For these reasons, I would like to humbly suggest that the "feed" link type is
reconsidered for inclusion; and that perhaps a better response to a perceived
lack of user uptake might be to better publicise its existence, rather than to
remove it from the spec.

Anyway, hope my thoughts on this are in some way constructive or useful -
please do feel very welcome to copy me into any further discussion of this
issue.

Best regards,
Ryan J. Bury
ryanjbury@gmail.com
www.rjbsoftware.co.uk

Posted from: 78.148.149.187