[wp-hackers] feed:http://address problem

[Owen Winkler]

Denis de Bernardy wrote:
> I think feed: takes us backwards, rather than forward. Comparing feed: to
> DOS file extensions would probably make sense. It is redundant.

It's only redundant if you're talking about using mime types to 
differentiate a feed from web content, and as I've pointed out, this 
doesn't work because the browser would forward the content of the page 
to the feed reader, not the URL of the feed.  This isn't useful.

Telling the browser to open a link with the feed: scheme means without 
doubt that the referenced URL is to be opened with a feed reader, and 
includes the required information to make it work.

> And usability-wise, it would make more sense if the browser
> handled feeds automatically, without any need for feed:. When you open a
> Word document in IE, IE asks if you want to save it or if you want to open
> Word to read it. Likewise, IE and FF should go: "Hey, my user just opened an
> RSS feed; I'll ask my user if he wants to subscribe to it or if he wants the
> raw XML."

Since a web site may wish to render any XML as a web page in a 
requesting browser, I don't see any way to automatically detect whether 
XML is a feed or a web page other than using the feed: scheme.  I would 
be really annoyed if my browser starts asking me whether to subscribe or 
view every web page that is written in XML, thinking that it might be a 
feed.

> Thus, and on closer inspection, I find myself wondering if MS will implement
> feed: at all, in IE7 and Longhorn. Surely, some MS usability person will
> pinpoint the above two points. They are strong arguments against. And
> possible show stoppers for widespread feed: adoption.

It's unlikely that MS will ever implement feed: in IE, since protocol 
handlers are built into the OS, and so when a request for a feed: URI is 
  handled, it is passed of to the registered application for that 
protocol.  (Windows calls these protocol handlers, even though they're 
not protocols but schemes.)  If your email client wasn't registered to 
the mailto: protocol, your browser would churn wondering what to do with 
links that point to it.

Since all feed readers that implement feed: in Windows are accomplishing 
it by registering their app for that protocol (and if they're not, they 
should because it's uber-easy), there really isn't a reason for MS to 
adopt the feed: protocol directly into IE - and this also doesn't imply 
that their lack of support for it implies a general lack of support.

Owen