On May 25, 2009, at 5:54 AM, Sam Ruby wrote:
[snip brisk disagreement over meta-issue]
(I'll note that I have no idea which contradictions you are
concerned with, perhaps you can point me to the relevant issues in
the issue tracker or bugzilla.)
The current draft contains content sniffing for feeds. If
accurately describes uniform browser behavior. It reinterprets
HTTP. It is not part of a vocabulary. It is not part of associated
API.
If it weren't contentious, it wouldn't be an issue. It is
contentious. One way to address it is to remove the section.
Another is to label it properly.
At first glance, it's not clear to me the detailed feed sniffing
algorithm needs to be in the spec. Sniffing feeds in generic XML types
is not described at all (presumably left to UA discretion), it's not
clear to me why sniffing feeds in HTML needs to be described in
detail. I don't think either sniffing feeds in XML or sniffing feeds
in HTML is essential to HTML UA interoperability. But perhaps this is
better discussed on public-html.
Indeed, sniffing of feeds from both XML types and text/html is not
browser-specific or HTML-specific. For example, NetNewsWire will
happily process the W3C RSS feed and the intertwingly.net Atom feed
when they are served as text/html. This may indicate that sniffing is
better described in a separate spec that is independent of HTML.
Removing accurate, but incomplete, labels does not address the issue.
I don't think replacing one accurate, but incomplete label with
another would address the issue either. I hope you can agree with that.
Regards,
Maciej
