[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [syndication] My take on shared feed lists
- To: <syndication@yahoogroups.com>
- Subject: Re: [syndication] My take on shared feed lists
- From: "Bill Kearney" <wkearney@syndic8.com>
- Date: Wed, 15 Oct 2003 10:31:21 -0400
- References: <3F8C650B.30608@wingedpig.com> <050d01c392c5$db06bb90$200ca8c0@wkearney.com> <20031015060618.GE4865@thermal> <059501c39321$67c28130$200ca8c0@wkearney.com> <B4fVoD3jPVj$EAf0@jblaptop.voidstar.com>
From: "Julian Bond" <julian_bond@voidstar.com>
> Bill Kearney <wkearney@syndic8.com> wrote:
> >http://somesite.hosted.example.com/users/jwz/weblog/default.xml
> >http://somesite.hosted.example.com/users/jwz/default.xml
> >http://somesite.hosted.example.com/users/default.xml
> >http://somesite.hosted.example.com/default.xml
>
> I can imagine the index.html page at each level having a <link> to a
> feedlist file that was different for each level. And the different
> feedlist files would probably contain links to some of each other.
Exactly. This as opposed to encouraging the bad idea of blindly trolling around
for a fixed filename. Better to let the pages speak for themselves and point to
whatever they need. Better to encourage the idea of head section parsing to get
them into the habit.
> Sorry. I just swapped from the discovery problem to the file content
> problem.
They're sort of related. In that the discovery format has a chance to help the
process along by stating what format is to be expected at the target URL.
Granted, people can screw this up. But one thing it /will/ help is the lack of
outgoing MIME types found on many sites. Many, many sites do not currently
output a proper MIME type for their existing RSS and XML data. Often they serve
it up as text/plain or text/xml. That's of little help regardless of any
filename extensions.
Using a type specifier gives the calling program a leg up on pulling only what
it needs in a format it understands. So if your page has upteem different
formats indicated my program can focus in on just the ones it knows it can
handle. Now, my program will have to make sure it's gotten actually useful data
but that's true regardless.
So basically if we promote the idea of link tags we get them used to the idea of
finding this and possibly more data there. We likewise promote the idea that
multiple formats might be possible and that their type should be specified. We
end up being able to support growth and also eliminate blindly wasting
bandwidth. It's a win-win all around.
-Bill Kearney