RE: [syndication] shared feed lists

["Roger B." <rogben@swbell.net>]

> Now please explain why is the first approach more efficient.

Dave,

I doubt it will *always* be more efficient, but in many cases, the
aggregator has already fetched the HTML to search for standard RSS
autodiscovery markup. Looking for a subscription link while you're there
seems logical. And if you don't find one, it makes sense to me to *then*
fall back to looking for a default file.

And again, a single, fixed-location file just doesn't work in some
environments. I don't generate static files, for starters. Then there's the
fact that freebie accounts don't get their own domains... they just get
subdirectories. And since any user can generate custom feeds that the parent
app doesn't know about, I can't just stockpile a master list somewhere.

For example, search-derived feeds...
http://makeashorterlink.com/?X1B161536 brings up a feed containing all of my
blog's RSS-Data posts. There's no way for Roger the administrator to know
that Roger the blogger wants to give visitors easy access to that particular
feed. In order for him to get any value out of myPublicFeeds, Roger the
blogger needs to be able to decide for himself what he wants to expose as a
collection.

(And while it's an edge case, there's another scenario that would lead to
confusion. If a user chooses not to publish a feed at all by blanking out
her default RSS template, there's no straightforward, programmatic way for
me to know about it. So even if I had the static file at the webroot with a
master list of all feeds, the user might end up with an empty document
rather than RSS. Sounds crazy I know, but I've seen people who *really*
don't like the idea of syndication.)

--
Roger Benningfield