Following:-
http://diveintomark.org/archives/2002/05/30.html#rss_autodiscovery

I just added the <link tag to http://www.ecademy.com/ and
http://www.voidstar.com/

What if there are two feeds? Do you think it's going to screw the
readers of the <link tag to just put two entries in?

For instance, Ecademy has a feed for the weblogs and another feed for
the articles. I've added both with the blogs one first.

--
Julian Bond Email&MSM: julian.bond@...
Webmaster:              http://www.ecademy.com/
Personal WebLog:       http://www.voidstar.com/
CV/Resume:          http://www.voidstar.com/cv/
M: +44 (0)77 5907 2173   T: +44 (0)192 0412 433

>What if there are two feeds? Do you think it's going to screw the
>readers of the <link tag to just put two entries in?

I wanted to ask this question as soon as I heard of this new momentum.
There are cases where a page could have any number of different feeds
associated with it.  And not just pages like the home page for Syndic8.com
either.

There's also the question of what about other applications that would like
to use text/xml?  Wasn't there discussion about using application/xml+rss or
something?

-Bill Kearney

_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp.

> What if there are two feeds? Do you think it's going to screw the
> readers of the <link tag to just put two entries in?
>
> For instance, Ecademy has a feed for the weblogs and another feed for
> the articles. I've added both with the blogs one first.

     I would think having the <link ... /> tag point to a feed directory for
multiple feeds would be better, maybe an OPML document. For only 2 feeds,
that is a bit obtuse, but in general, I think a directory would be better.

James

>     I would think having the <link ... /> tag point to a feed directory for
> multiple feeds would be better, maybe an OPML document. For only 2 feeds,
> that is a bit obtuse, but in general, I think a directory would be better.

This would be hard for a scriptlet to handle, since it would have to
fetch a separate document into a hidden window or frame. It would
be easier for a scriptlet to extract *all* of the relevant link tags and
then pass those on to the underlying site as an argument list. The
site can then present a web-based UI to allow the user to choose
any or all of the feeds.

For example, here's a scriptlet that I found:

oid(d=document);
void(el=d.getElementsByTagName('link'));
for(i=0;i<el.length;i++)
{
   if(el[i].getAttribute('rel').indexOf('alternate') != -1 &&
el[i].getAttribute('type').indexOf('text/xml')!=-1)
   {
    
void(location.href='http://www.syndic8.com/feedinfo.php?FeedDataURL='+el[i].getA\
ttribute('href'))
   };
};

It would not be hard to modify it to accumulate all of the tags and then pass
all of them to the site.

Jeff;

----- Original Message -----
From: "James Linden" <jlinden@...>
To: <syndication@yahoogroups.com>
Sent: Saturday, June 01, 2002 12:59 AM
Subject: Re: [syndication] RSS Discovery


> > What if there are two feeds? Do you think it's going to screw the
> > readers of the <link tag to just put two entries in?
> >
> > For instance, Ecademy has a feed for the weblogs and another feed for
> > the articles. I've added both with the blogs one first.
>
>     I would think having the <link ... /> tag point to a feed directory for
> multiple feeds would be better, maybe an OPML document. For only 2 feeds,
> that is a bit obtuse, but in general, I think a directory would be better.
>
> James
>
>
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>

On Sat, 1 Jun 2002, Dan Kohn wrote:

> It should be application/rss+xml, as text/xml may very well be
> deprecated in the next version of RFC 3023.
>
> http://www.ietf.org/internet-drafts/draft-nottingham-rss-media-type-00.t
> xt
>
> See RFC 3023 for the reasons text/xml is a bad idea (especially the
> charset and display issues).  Could you please forward this to
> appropriate parties and/or lists?

Makes sense to me. cc: list inflated to include rss-dev and syndication;
any extended followup should probably migrate from fork to those lists.

cheers,

dan

>
>           - dan
> --
> Dan Kohn <mailto:dan@...>
> <http://www.dankohn.com/>  <tel:+1-650-327-2600>
> Essays announced on <mailto:dankohn-subscribe@yahoogroups.com>
>
> -----Original Message-----
> From: Dan Brickley [mailto:danbri@...]
> Sent: Saturday, June 01, 2002 11:41
> To: Gary Lawrence Murphy
> Cc: fork
> Subject: Re: Automatic discovery of RSS
>
>
> On 1 Jun 2002, Gary Lawrence Murphy wrote:
>
> >
> > Just in case it got missed, this use of the link tag is such an
> > obvious and wonderful idea, it's no wonder the complete re-tooling of
> > the entire blog space will probably happen in a matter of days;
> > certainly our site RDF was link-enabled within seconds of reding this:
> >
> >     http://www.oreillynet.com/cs/weblog/view/wlg/1475
> >
> >     Two days ago, Matt Griffith has come up with the smart idea of
> >     using the HTML link element to point to a site's RSS feed from the
> >     site itself - thus allowing automatic discovery of RSS feeds.
> >
> >     By adding a line like this: <link rel="alternate" type="text/xml"
> >     title="XML" href="http://rss.benhammersley.com/index.rss" /> a
> >     site would be providing metadata as to the location of its feed -
> >     and this would allow newsreaders, browsers and search engines to
> >     automatically locate the feed.
>
> Using HTML LINK is nifty. I think this was discussed a while back on
> RSS-DEV but ran out of energy as nobody was sure which mime-type to use.
> Pick one and run with it seems a good plan. I'm not sure text/xml is
> quite
> specific enough, but this is something that shouldn't slow down
> experimental deployment. Beats having central repositories...
>
> Dan
>
>
>
> http://xent.com/mailman/listinfo/fork
>

Hi

On Sat 01-Jun-2002 at 03:14:53PM -0400, Dan Brickley wrote:
>
> > Using HTML LINK is nifty. I think this was discussed a while back on
> > RSS-DEV but ran out of energy as nobody was sure which mime-type to
> > use.

It was also discussed on www-rdf-interest:

   http://lists.w3.org/Archives/Public/www-rdf-interest/2002Jan/0074.html

Personally I have been doing this for a while:

   <link
     rel="syndication"
     href="rss100headline.rdf"
     type="application/xml"
     hreflang="en-gb" />

   <link
     rel="alternate syndication"
     href="rss090headline.rdf"
     type="application/xml"
     hreflang="en-gb" />

Chris

--
Chris Croome                               <chris@...>
web design                             http://www.webarchitects.co.uk/
web content management                               http://mkdoc.com/
everything else                               http://chris.croome.net/

I think this is one of those questions where the answer depends on who you
talk to. :)

The "Internet" way of doing it would be to use media types (aka MIME
types). This would identify each format/language with a media type, like
application/rss+xml. The problem with this is that one isn't yet
registered for RSS, and the process for registering new types that aren't
in a vendor-specific namespace is lengthy. (as an aside, I wonder of one
of the namespace cooperatives like xmlns or purl would consider being the
root of a vnd tree for media types, so that community formats like RSS1.0
could avoid the pain).

The "HTML" way of doing it is to use a new rel="...", as you've done. The
problem is that there's no way to manage that namespace, and this
mechanism doesn't seem to have caught on very well for extensions.

The "Web" way of doing it is with a URI, but I haven't seen a way to
identify a link format in a <LINK>. Or am I missing it?

The most practical and probably the most correct way to do this is with a
media type, especially in light of the TAG's recent findings. Now, if
RSS-DEV would *AHEM* go ahead and submit it as a note, we'll be set.

Cheers,



----- Original Message -----
From: "Chris Croome" <chris@...>
To: <rss-dev@yahoogroups.com>
Cc: "Dan Kohn" <dan@...>; "Gary Lawrence Murphy"
<garym@...>; "fork" <fork@...>; <syndication@yahoogroups.com>
Sent: Tuesday, June 11, 2002 8:51 AM
Subject: [syndication] Re: [RSS-DEV] RE: Automatic discovery of RSS


> Hi
>
> On Sat 01-Jun-2002 at 03:14:53PM -0400, Dan Brickley wrote:
> >
> > > Using HTML LINK is nifty. I think this was discussed a while back on
> > > RSS-DEV but ran out of energy as nobody was sure which mime-type to
> > > use.
>
> It was also discussed on www-rdf-interest:
>
>   http://lists.w3.org/Archives/Public/www-rdf-interest/2002Jan/0074.html
>
> Personally I have been doing this for a while:
>
>   <link
>     rel="syndication"
>     href="rss100headline.rdf"
>     type="application/xml"
>     hreflang="en-gb" />
>
>   <link
>     rel="alternate syndication"
>     href="rss090headline.rdf"
>     type="application/xml"
>     hreflang="en-gb" />
>
> Chris
>
> --
> Chris Croome                               <chris@...>
> web design                             http://www.webarchitects.co.uk/
> web content management                               http://mkdoc.com/
> everything else                               http://chris.croome.net/
>
>
>
> Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/
>
>

>Yesterday, my RSS feed got more hits than the home page, and my RSS feed has
  >been up for less than a week. Home page hits are off significantly, which
  >tells me that a number of people who used to hit the site several times a
  >day (you wonderful people) have plugged my site into their RSS aggregator.
  >Granted, depending on the feed reader you use, you might be hitting the site
  >50 times a day with it, so there's no way to tell how many people are
  >actually using the RSS feed (well, there is, but not with the tool I'm using
  >for log analysis), but it does seem to have taken off like a rocket.
  >http://www.rc3.org/cgi-bin/less.pl?arg=4156


--
Morbus Iff ( i'm the droid you're looking for )
Culture: http://www.disobey.com/ and http://www.gamegrene.com/
Tech: http://www.oreillynet.com/pub/au/779 - articles and weblog
icq: 2927491 / aim: akaMorbus / yahoo: morbus_iff / jabber.org: morbus

I've been fooling around with the site images collected from the
various RSS feeds at Syndic8 and started to composite them into
some larger pictures:

http://www.syndic8.com/SiteImageTest/

Click anywhere to see a larger panel.

The results are quite interesting:

1. 88x31 is a nice recommendation, but there's a lot of images that
    don't fit.

2. A lot of sites don't bother with images yet (I don't have the stats
     for this at hand, but trust me on this).

3. This is going to be a terrific way to navigate through "feedspace";
     stay tuned for the next obvious step in this direction.

4. Too many sites use default images.

If a large number of feeds reference the same image URL (e.g.
Moreover or NewsIsFree) then a single image is rendered.

This is just for fun and for feasability testing, but I thought it was
kind of cool and wanted to share it.

Jeff;


Jeff Barr - Vertex Development - jeff@...
   IM:               MSN: jeffscottbarr@...; AIM: jeffscottbarr
   Resume:       www.vertexdev.com/~jeff/resume.html
   Good Stuff:   www.syndic8.com   www.headlineviewer.com

>  >Granted, depending on the feed reader you use, you might be hitting the site
>  >50 times a day with it, so there's no way to tell how many people are
>  >actually using the RSS feed (well, there is, but not with the tool I'm using
>  >for log analysis), but it does seem to have taken off like a rocket.
>  >http://www.rc3.org/cgi-bin/less.pl?arg=4156

There's apparently been a recent change to Radio that will allow it to alter
it's user-agent to reflect account info.  That way the log would truly show
who's been loading the RSS feed.  It wouldn't tell you what content they'd been
reading however.  You'd have to do some funky stuff with redirects and custom
<link> elements to make that happen.

-Bill

> http://www.syndic8.com/SiteImageTest/
>
> This is just for fun and for feasability testing, but I thought it was
> kind of cool and wanted to share it.

Can you not display the larger images?  I'd hate to see folks think they should
start using bigger pictures in order to stand out on a page like this.  Either
that or force-fit them into a smaller box by scaling them down?  Perhaps
distorting them as a way to 'encourage' the to start using the proper image
sizes?

-Bill Kearney

My generator program uses a threshold to stop displaying
images larger than a certain size. Right now it is set to
show anything up to 300x300.

Jeff;

----- Original Message -----
From: "Bill Kearney" <wkearney99@...>
To: <syndication@yahoogroups.com>
Sent: Wednesday, June 12, 2002 3:32 PM
Subject: Re: [syndication] The wonderful world of site images....


> > http://www.syndic8.com/SiteImageTest/
> >
> > This is just for fun and for feasability testing, but I thought it was
> > kind of cool and wanted to share it.
>
> Can you not display the larger images?  I'd hate to see folks think they
should
> start using bigger pictures in order to stand out on a page like this.  Either
> that or force-fit them into a smaller box by scaling them down?  Perhaps
> distorting them as a way to 'encourage' the to start using the proper image
> sizes?
>
> -Bill Kearney
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>

>There's apparently been a recent change to Radio that will allow it to alter
>it's user-agent to reflect account info.  That way the log would truly show
>who's been loading the RSS feed.  It wouldn't tell you what content they'd
>been
>reading however.  You'd have to do some funky stuff with redirects and custom
><link> elements to make that happen.

Yup, that is also possible under the new AmphetaDesk (the Referer thing).

--
Morbus Iff ( HOW DO I DELIT TEH TREE FILEZ?!@! )
Culture: http://www.disobey.com/ and http://www.gamegrene.com/
Tech: http://www.oreillynet.com/pub/au/779 - articles and weblog
icq: 2927491 / aim: akaMorbus / yahoo: morbus_iff / jabber.org: morbus

It seems most of the RSS paradigm leans towards "broadcasting," ie,
the audience of the RSS item is as many folks as possible. I've been
thinking about how useful it might be to make "saved searches" RSS
documents. A user of my search engine could then add the RSS document
of a saved search to the rest of their collection of RSS documents.

The problem is the very personalized nature of the feed. My
understanding of how most RSS aggregators work is that a central
"authority" identifies and adds feeds to a particular interface. Thus
I can't add my "saved search on foo" so that only I can see it. I can
imagine that there's an existence proof that some interface out there
allows the ueser to directly add their own feeds, but what is desired
is that a user of my search engine could add their "saved search on
foo" (and *not* their "saved search on bar") to their favorite
aggregator.

Am I right in my understanding of this problem? I have to admit to a
merely theoretical understanding of RSS with the belief that I
probably use it more often than I know (except for early experiments
with my.netscape).

Thanks,

judith

PS: I was happy to discover, I'm not the only person to want saved
searches available ia RSS: "In the future, I hope to retrieve saved
searches from the popular news engines like World News via RSS feeds
and postings to listservs to which I subscribe. " See
http://www.llrx.com/features/rssforlibrarians.htm

Very much so. I'd think that this is a good marriage of REST and RSS; a
search engine might have a URI for a set of search results that looks like
   http://search.example.com/query?term=foo&term=bar
which returns text/html by default. To get the RSS feed, one GETs

http://search.example.com/query?term=foo&term=bar&type=application/rss+xml
or similar.

On a related note, one of the things that I'd like to see is RSS to be
offered as a more dynamic, customizable feed in other contexts as well.
For example. Slashdot allows users to customise what categories they see,
but the RSS feed is just a firehose of whatever's published. RSS is just
another format that's more machine-digestable than HTML; when the content
is list-based, it should be available in both representations.

Cheers,

----- Original Message -----
From: "judielaine" <jeb@...>
To: <syndication@yahoogroups.com>
Sent: Thursday, June 13, 2002 1:39 PM
Subject: [syndication] "Narrowcasting" RSS


> It seems most of the RSS paradigm leans towards "broadcasting," ie,
> the audience of the RSS item is as many folks as possible. I've been
> thinking about how useful it might be to make "saved searches" RSS
> documents. A user of my search engine could then add the RSS document
> of a saved search to the rest of their collection of RSS documents.
>
> The problem is the very personalized nature of the feed. My
> understanding of how most RSS aggregators work is that a central
> "authority" identifies and adds feeds to a particular interface. Thus
> I can't add my "saved search on foo" so that only I can see it. I can
> imagine that there's an existence proof that some interface out there
> allows the ueser to directly add their own feeds, but what is desired
> is that a user of my search engine could add their "saved search on
> foo" (and *not* their "saved search on bar") to their favorite
> aggregator.
>
> Am I right in my understanding of this problem? I have to admit to a
> merely theoretical understanding of RSS with the belief that I
> probably use it more often than I know (except for early experiments
> with my.netscape).
>
> Thanks,
>
> judith
>
> PS: I was happy to discover, I'm not the only person to want saved
> searches available ia RSS: "In the future, I hope to retrieve saved
> searches from the popular news engines like World News via RSS feeds
> and postings to listservs to which I subscribe. " See
> http://www.llrx.com/features/rssforlibrarians.htm
>
>
>
>
>
>
> Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/
>
>

> On a related note, one of the things that I'd like to see is RSS to be
> offered as a more dynamic, customizable feed in other contexts as well.
> For example. Slashdot allows users to customise what categories they see,
> but the RSS feed is just a firehose of whatever's published. RSS is just
> another format that's more machine-digestable than HTML; when the content
> is list-based, it should be available in both representations.

Search for Rael's work on Meerkat and it's parameter setups.  This is well
plowed soil here.

I'm in favor of parameterized RSS URLs being used to focus the content.  I'll
offer the opinion that attempting to use it as a means to "filter out" something
will undoubtedly end up losing something you'd rather not miss.  The
ability/reliability of people applying effective metadata (like categorization)
on items is still a very problematic area.  Fundamentally people don't generally
see the value to expending the effort.  Everyone wants categories but nobody can
decide which are truly relevant.  Even less are willing to expend the trivial
amount of energy to click even single checkboxes that would help others get a
grip on the content.  Of course, synthetic metadata is often worse...

-Bill Kearney

On Thu, 2002-06-13 at 13:54, Mark Nottingham wrote:
> On a related note, one of the things that I'd like to see is RSS to be
> offered as a more dynamic, customizable feed in other contexts as well.
> For example. Slashdot allows users to customise what categories they see,
> but the RSS feed is just a firehose of whatever's published. RSS is just
> another format that's more machine-digestable than HTML; when the content
> is list-based, it should be available in both representations.
Actually most links take content_type=rss as an option. So using this
and say the search engine you can pull quite a bit of data out of a
slash site (depends a little on what the site allows to be exported, for
instance Slashdot does not export descriptions).
Search, submissions, Journals, comments (via search), our rss feeds, and
quite a few other things are all available via RSS. For each Slash site
it is a matter of what has been enabled or not.

	 -Brian
--
_______________________________________________________
Brian "Krow" Aker, brian@...
Slashdot Senior Developer
Seattle, Washington
http://tangent.org/~brian/
http://askbrian.org/
_______________________________________________________
You can't grep a dead tree.

Agreed on all points. This is different, however; it's not categorizing
the feed, it's making the feed a first-class citizen, alongside the
"normal" content.


----- Original Message -----
From: "Bill Kearney" <wkearney99@...>
To: <syndication@yahoogroups.com>
Sent: Thursday, June 13, 2002 2:01 PM
Subject: Re: [syndication] "Narrowcasting" RSS


> > On a related note, one of the things that I'd like to see is RSS to be
> > offered as a more dynamic, customizable feed in other contexts as
well.
> > For example. Slashdot allows users to customise what categories they
see,
> > but the RSS feed is just a firehose of whatever's published. RSS is
just
> > another format that's more machine-digestable than HTML; when the
content
> > is list-based, it should be available in both representations.
>
> Search for Rael's work on Meerkat and it's parameter setups.  This is
well
> plowed soil here.
>
> I'm in favor of parameterized RSS URLs being used to focus the content.
I'll
> offer the opinion that attempting to use it as a means to "filter out"
something
> will undoubtedly end up losing something you'd rather not miss.  The
> ability/reliability of people applying effective metadata (like
categorization)
> on items is still a very problematic area.  Fundamentally people don't
generally
> see the value to expending the effort.  Everyone wants categories but
nobody can
> decide which are truly relevant.  Even less are willing to expend the
trivial
> amount of energy to click even single checkboxes that would help others
get a
> grip on the content.  Of course, synthetic metadata is often worse...
>
> -Bill Kearney
>
>
>
> Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/
>
>

That is so cool. Thanks!

Now, if we get the RSS discovery stuff all lined up, it'll eventually all
"just work" (so i don't have to find out like this ;)


----- Original Message -----
From: "Brian Aker" <brian@...>
To: <syndication@yahoogroups.com>
Sent: Thursday, June 13, 2002 1:57 PM
Subject: Re: [syndication] "Narrowcasting" RSS


> On Thu, 2002-06-13 at 13:54, Mark Nottingham wrote:
> > On a related note, one of the things that I'd like to see is RSS to be
> > offered as a more dynamic, customizable feed in other contexts as
well.
> > For example. Slashdot allows users to customise what categories they
see,
> > but the RSS feed is just a firehose of whatever's published. RSS is
just
> > another format that's more machine-digestable than HTML; when the
content
> > is list-based, it should be available in both representations.
> Actually most links take content_type=rss as an option. So using this
> and say the search engine you can pull quite a bit of data out of a
> slash site (depends a little on what the site allows to be exported, for
> instance Slashdot does not export descriptions).
> Search, submissions, Journals, comments (via search), our rss feeds, and
> quite a few other things are all available via RSS. For each Slash site
> it is a matter of what has been enabled or not.
>
> -Brian
> --
> _______________________________________________________
> Brian "Krow" Aker, brian@...
> Slashdot Senior Developer
> Seattle, Washington
> http://tangent.org/~brian/
> http://askbrian.org/
> _______________________________________________________
> You can't grep a dead tree.
>
>
>
>
> Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/
>
>

On Thu, 2002-06-13 at 14:09, Mark Nottingham wrote:
> That is so cool. Thanks!
I think so too. We will continue putting this in the code but will also be
releasing a formal SOAP plugin for Slash that does a lot more.
The design is done, just a matter of either myself or pudge writing it.


> Now, if we get the RSS discovery stuff all lined up, it'll eventually all
> "just work" (so i don't have to find out like this ;)
Finding RSS is still the hard part. OCS feeds are a pain to deal with
(and I think that it is dead now as a technology from what I can see).

Keep in mind that if you ever connect to web station running mod_mp3
under Apache it will spit out play lists and what is current playing in
RSS too  (a personal plug).

	 -Brian
--
_______________________________________________________
Brian "Krow" Aker, brian@...
Slashdot Senior Developer
Seattle, Washington
http://tangent.org/~brian/
http://askbrian.org/
_______________________________________________________
You can't grep a dead tree.

> The problem is the very personalized nature of the feed. My
> understanding of how most RSS aggregators work is that a central
> "authority" identifies and adds feeds to a particular interface.

No, a web portal site that's presenting the contents of RSS feeds may do that.
Some, like newsisfree, give you the ability to sign up and have it keep track of
your selected channels.  Others don't.  But a personal aggregator like peerkat,
blosxom, amphetadesk, headline viewer, feed reader, newzcrawler and others do
let you control which feeds are being shown.

> Thus
> I can't add my "saved search on foo" so that only I can see it. I can
> imagine that there's an existence proof that some interface out there
> allows the ueser to directly add their own feeds, but what is desired
> is that a user of my search engine could add their "saved search on
> foo" (and *not* their "saved search on bar") to their favorite
> aggregator.

Sure and using a personal aggregator lets you do this.

> Am I right in my understanding of this problem? I have to admit to a
> merely theoretical understanding of RSS with the belief that I
> probably use it more often than I know (except for early experiments
> with my.netscape).

my.netscape was indeed a web portal that allowed very little per-user control.
Fortunately things have changed since then!

> PS: I was happy to discover, I'm not the only person to want saved
> searches available ia RSS: "In the future, I hope to retrieve saved
> searches from the popular news engines like World News via RSS feeds
> and postings to listservs to which I subscribe. " See
> http://www.llrx.com/features/rssforlibrarians.htm

Mailing lists running up on yahoo are already capable of providing an RSS feed
of new items as they're posted:
http://www.syndic8.com/feedlist.php?ShowMatch=groups.yahoo.com

There's cool stuff that can be accomplished with syndication.  I'd argue that
not everything is suitable for it.  All these folks running around with
aggregators, outliners and portals might do well to hold up a second and look at
the big picture.  Syndication stands to play a much larger part than it has in
the past.  But there's still plenty of situations where good old web pages and
thick clients still excel.

-Bill Kearney

Bill Kearney said: [1]

> [RSS 1.0 is] designed to require a feed to contain it's own source URL.

I'm coming to this party late, so I had to read the spec. This is too hazy
to be a requirement:

<blockquote>The {resource} URL of the channel element's rdf:about attribute
must be unique with respect to any other rdf:about attributes in the RSS
document and is a URI which identifies the channel. <strong>Most
commonly</strong>, this is either the URL of the homepage being described
<strong>or</strong> a URL where the RSS file can be found. [emphasis mine]
</blockquote>

We may want GUIDs to identify the feeds uniquely, but while we're fetching
them from a URL, it would be nice [2] for subscribers to that URL to update
it when the URI changes.

The server could notify you using HTTP redirects, but, as has been
mentioned, that fails when HTTP is not the transport. It fails even when
HTTP _is_ the transport because aggregators don't update in either case.

Take my site, for example. I moved. I've permanently redirected the URL.
People are still reading the feed, in its current state, because of the
redirect. And Syndic8 has me, kindly enough, at both URLs [3] -- because
Syndic8 doesn't notice the permanent redirect.

Google doesn't either. My old site is still more authoritative than the new
one, perhaps because of the sticky links from other pages to it.

But when I cancel my account on that server, the subscriptions -- and
Google's index -- will fail.

/cwc

[1] http://rss.benhammersley.com/archives/000054.html#000054
[2] http://www.coxesroost.net/peanuts/2002/05/10.html#a365
[3]
http://www.syndic8.com/feedlist.php?PHPSESSID=ec7be5bfa081a918af4806ffef107f
c7&ShowMatch=peanut+gallery&ShowStatus=all

--

Smart networks are a Layer 8 problem.
http://www.coxesroost.net/peanuts/

The new version of AmphetaDesk is available. You can grab it from:

   http://www.disobey.com/amphetadesk/

And can find more information about this release here:

   http://www.disobey.com/amphetadesk/whats_new.html

Let me know what you think, and if you mention it somewhere, eh? There's
also an "in the know" random interview with yours truly available below. It
covers some of the features / mentality behind the newest AmphetaDesk:

    http://blogspace.com/swhack/weblog/2002/06/13/#i1024002651.914514

--
Morbus Iff ( strive for mediocrity )
Culture: http://www.disobey.com/ and http://www.gamegrene.com/
Tech: http://www.oreillynet.com/pub/au/779 - articles and weblog
icq: 2927491 / aim: akaMorbus / yahoo: morbus_iff / jabber.org: morbus

In article <aeavu0+saoc@eGroups.com>, judielaine <jeb@...>
writes
>It seems most of the RSS paradigm leans towards "broadcasting," ie,
>the audience of the RSS item is as many folks as possible. I've been
>thinking about how useful it might be to make "saved searches" RSS
>documents. A user of my search engine could then add the RSS document
>of a saved search to the rest of their collection of RSS documents.

Maybe I've misunderstood but there are several search sites out there
that can optionally return results as RSS. Moreover, Snewp, NWFusion
spring to mind. I'm reading synthetic feeds from search parameters from
all three of these. Both in a personal aggregator and centralized ones
built with Drupal.

You say "a user of my search engine". If you really mean "My" then it's
a simple matter of code to output RSS instead of html and then letting
the users know it's available.

>The problem is the very personalized nature of the feed. My
>understanding of how most RSS aggregators work is that a central
>"authority" identifies and adds feeds to a particular interface.

That's just a matter of design. There are plenty of personal
aggregators, as well as central aggregators that allow personal
customization.

>PS: I was happy to discover, I'm not the only person to want saved
>searches available ia RSS: "In the future, I hope to retrieve saved
>searches from the popular news engines like World News via RSS feeds
>and postings to listservs to which I subscribe. " See
>http://www.llrx.com/features/rssforlibrarians.htm

I would love to see Google optionally return RSS from their normal
search engine. But then I'd also love to see Yahoogroups (and
livejournal) return the first 100 chars of the message in the
<description>. And there's a bunch of mainstream journalism sites that
I'd love to see support RSS. And the big money-big profile CMS don't yet
seem to output RSS by default. And the NYTimes really ought to have an
RSS feed instead of a proprietary format.

Even though Syndic8 now has >5000 feeds recorded, there's still more
evangelism to be done.

--
Julian Bond Email&MSM: julian.bond@...
Webmaster:              http://www.ecademy.com/
Personal WebLog:       http://www.voidstar.com/
CV/Resume:          http://www.voidstar.com/cv/
M: +44 (0)77 5907 2173   T: +44 (0)192 0412 433

There's been some talk on another list and also on a couple of blogs[1]
about aggregating RSS feeds together and then using it to create a new
synthetic feed.

Inspired by this, I added a bit of functionality to Ecademy's news
system[3]. We already collect a 100 or so news feeds and group them into
topics based on the feed source[4]. Drupal's news system already had all
this functionality where by a feed can be given one or more attributes
which are taken from a list of topics each of which has one or more
attributes. The topic display groups news items according to the
attribute matches.

The addition was simple. Just take a topic and write the RSS back out
again for the items in that topic. So now we've got a composite feed
available on (say) "Web Services" made up of 20 source feeds[5].

Every so often we get to talking about categorization of items and feeds
within the standard. This approach of getting the readers to do the
categorization and pushing back out composite feeds might be a nicely
low tech and distributed way of getting some useful grouping done.

[1] http://www.looselycoupled.com/blog/2002_06_09_lc.htm#85164156
[3] http://www.ecademy.com/module.php?mod=import
[4] http://www.ecademy.com/module.php?mod=import&op=bundles
[5] http://www.ecademy.com/module.php?mod=import&op=rssbundle&id=12

--
Julian Bond Email&MSM: julian.bond@...
Webmaster:              http://www.ecademy.com/
Personal WebLog:       http://www.voidstar.com/
CV/Resume:          http://www.voidstar.com/cv/
M: +44 (0)77 5907 2173   T: +44 (0)192 0412 433

Hi all,

OCS[1] has been at version 0.4 for over a year and it's time to revisit
it. One of the major flaws of OCS has been that it is not strictly
correct RDF which means that implementors can't use an RDF parser to
process it and standard xml methods are complicated by the nearly-RDF
syntax.

I'd like to take this opportunity to take a poll[2] of opinion on the
future of RDF in OCS. (This poll will close at noon (GMT) 30 June
2002)

Thanks,

Ian

[1] http://internetalchemy.org/ocs/
[2] http://pub.alxnet.com/poll?id=2286452

> Take my site, for example. I moved. I've permanently redirected the URL.
> People are still reading the feed, in its current state, because of the
> redirect. And Syndic8 has me, kindly enough, at both URLs [3] -- because
> Syndic8 doesn't notice the permanent redirect.

There are two things to consider.  One, syndic8 DOES notice when legitimate
redirects are used and attempts to note them.  Two, keeping the old URL serves
as a marker for anything that has the old URL.  That way anything looking up the
old one will be able to find a reference to the new database record.  This will
allow Syndic8 to do stuff like provide intelligent redirection, perhaps in the
form of a side-band channel of instructions.

I personally have put a fair effort into helping other sources of multiple feeds
understand how to use more specific error codes like 410 instead of just 403 and
the like.  Just using the existing HTTP error code standards would help
*enormously*.

So it's not like we didn't think about it and do nothing.

-Bill Kearney

Hi all,

Can folks chime in with opinion as to what's supposed to be used in the dc:date
element?

The specs for RSS-0.91 are ambiguous about using pubDate and lastBuildDate.  I'd
like to get a clarification as to how dates are to be indicated in an RSS-1.0
feed.

I can see several issues.  One is the timestamp of when the XML was truly
generated.  Some are statically generated so this timestamp could be considered
some way to check whether a local copy is any 'different' than the one online.
However, a great many feeds are dynamically generated making this timestamp
completely unsuitable for use as a checking indicator.  I'm not unaware of
e-tags and content-header stuff, let's focus on in-feed timestamping for now.  I
recognize protocol negotiation is an important tool it's just outside the scope
of what I'm clarifying here.

Another issue concerns when the content of the items has changed.  A timestamp
indicating when the most recent item within a feed has changed would be quite
helpful for more environments.  This raises the question of whether the
timestamp of content change is item-related or whether something about the
channel as a whole has been altered.  I'd suggest that most folks only really
care about a content timestamp change but the header question does deserve some
consideration.  That way they could tell if the header-related material for the
channel has also changed.

Additionally, how about discussing a dc:date inside the items themselves?  Most
environments DO have the ability to extract this information and a great many
users would take advantage of it.  Can we speak about it?  The module for
aggregation does support it.

The recent dc_terms module also suggests created and modified date, but the
examples only suggest it's use in the header.

I'd like to get a consensus going on making more effective use of dates inside
the channel header and inside items themselves.

-Bill Kearney

Crossposted to syndication and syndic8.

Hey all. John S. Rhodes just released an interview with me which covers a
bit of AmphetaDesk, a bit of syndication in general (mentioning Radio
Userland, Syndic8, HotSheet, etc.) and a bit of RSS and XML in plain text
speak for the new user. Go check it out. I think it came out rather well.

  http://webword.com/interviews/iff.html

--
Morbus Iff ( united we're bland )
Culture: http://www.disobey.com/ and http://www.gamegrene.com/
Tech: http://www.oreillynet.com/pub/au/779 - articles and weblog
icq: 2927491 / aim: akaMorbus / yahoo: morbus_iff / jabber.org: morbus

Thanks for the link!
I am totally new to this schtuff so if anyone knows of a site that speaks
to the layman, please send me a link so I don't harass YOU too much ;-)

At 10:50 AM 6/18/2002 -0400, you wrote:

>Crossposted to syndication and syndic8.
>
>Hey all. John S. Rhodes just released an interview with me which covers a
>bit of AmphetaDesk, a bit of syndication in general (mentioning Radio
>Userland, Syndic8, HotSheet, etc.) and a bit of RSS and XML in plain text
>speak for the new user. Go check it out. I think it came out rather well.
>
>  http://webword.com/interviews/iff.html
>
>--
>Morbus Iff ( united we're bland )
>Culture: http://www.disobey.com/ and http://www.gamegrene.com/
>Tech: http://www.oreillynet.com/pub/au/779 - articles and weblog
>icq: 2927491 / aim: akaMorbus / yahoo: morbus_iff / jabber.org: morbus
>
>
>
>Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/