Re: [syndication] Re: shared feed lists

["Bill Kearney" <ml_yahoo@ideaspace.net>]

Good points.

Things like DNS are capable of doing this now, but for host or domain-wide
resources.  One could, conceivably, setup a DNS record that could be resolved to
a hostname that had the default URL.  That'd be sort of overkill though.  I'm
not sure the use of TEXT records is the ideal way to go about this though.
There's ways to use DNS to indicate geo-positioning also.  But as you rightly
point out this is even /less likely/ to be something most sites can, let alone
would, set up.

But things like SMTP and IM can use this quite effectively.  user@domain
resolves back to smtp.domain.example.com or im_gateway.domain.example.com.
There's a lot to DNS that's more than just hostname to IP address lookups.

Right now Syndic8 publishes my personal lists idea.  We can easily implement any
new format for this as it's largely template driven.  I'd raised the question on
aggregators as to how readers would react to new namespaced elements being used
in OPML files.  Intial feedback seems to indicate no readers had trouble seeing
past the added data.

As to ADX, it's as useful as anything else I suppose, here's a list in it's
format:
http://www.syndic8.com/genfeed.php?Format=adx&UserID=wkearney99&List=8

Got a few rough spots, all of which are easily fixed if sufficient interest
develops.  We don't track a per-user language setting; only per-feed and the
logic for checking lets a blank one slip past.   So we can't show "your group"
as being in a given language but we can show you the per-feed languages.  I'm
thinking as an interim step this is useful enough for most purposes.  But having
an index-wide language tag is a reasonable request.

I'm also not entirely clear on what the IndexRef is supposed to be pointing
toward.   Is it pointing to itself?  Or to another file?  Sort of like
rdfd:seeAlso?  If so, it'd be helpful to allow use of a type attribute of some
kind to let the consuming program know what format the external document is
using.  Or is this expecting to be an ADX-only world?  (not that there's
anything wrong with that idea).

Anyone with a personal list can get it as ADX, just change the 'Format=opml' to
'Format=adx'.  Bearing in mind, however, that this is not an 'officially
supported' format and the URL may break without notice.

-Bill Kearney


----- Original Message -----
From: "tappnel" <app@mplode.com>
To: <syndication@yahoogroups.com>
Sent: Tuesday, October 14, 2003 1:05 PM
Subject: [syndication] Re: shared feed lists


> For the sake of discussion here is something related that James Snell wrote up
on
> auto-discovery of feeds and other services[1]. It stems from Jeremy's original
post on
> RSS autodiscovery a month ago. I commented earlier today here.[2] In short,
James
> put up a rough draft of a format he is calling Automatic Discovery indeX
(ADX). In
> short, ADX is stright-forward, inclusive to other formats and services, can be
> validated and builds on the prior art of WSIL.
>
> Intelligent and constructive feedback welcome.
>
> As for the debate to how to find auto-discovery I'm somewhat on the fence. I
just
> assume use index.* for the file name. I think as Jeremy suggests will have to
support
> multiple forms. For instance in James' piece he suggests using DNS Service
> Discovery. I'm personally not sure about this method because most users will
not be
> able to do this for one reason or another. However it may be quite attractive
to some
> like Yahoo! who clearly has the access and expertise. Yet another method of
> discoverying feeds or a feeds list that has not been discussed is through a
public
> directory service like Syndic8 or other.
>
> <tim/>
>
> [1] http://www.snellspace.com/blog/archivesNew/000065.html
> [2] http://www.mplode.com/tima/archives/000321.html