Re: [syndication] Re: SuperOpenRatings

["Dave Winer" <dave@userland.com>]

Sorry for the confusion, it's been a busy day, and this thread began on
another mail list.

There are publishers of original content that change the URL even though the
story they're pointing to is not new.

I don't know why they do it, but some do.

I don't consider the canonical <item> id to be a high priority, but Aaron
suggested it (I think, he'll correct me if I remember it incorrectly) and I
asked him to post the suggestion here.

Dave


----- Original Message -----
From: "Robert A. Lentz" <ralentz@ralentz.com>
To: <syndication@yahoogroups.com>
Sent: Monday, February 05, 2001 5:53 PM
Subject: Re: [syndication] Re: SuperOpenRatings


> At 08:00 PM 2/5/01, you wrote:
> >The idea is to deal with people who have CMSes that change the link as
the
> >document is moved.
>
> (Okay, this might be slightly off topic...)
>
> Please forgive my naivete, but if I publish a syndication file pointing to
> new articles on my web site, there are CMSes out there which then change
> the URL I provided that points to those items? Sounds like they are broken
> to me.
>
> >Combine this with misspelled / reused / missing titles
>
> Then use the unique link.
>
> Aren't we talking about third parties (customers) rating content generated
> by other people?
>
> >and the fact that the link is option in RSS 0.92 and you've got a
problem.
>
> I guess this is what I get for ignoring 0.92 until it is finalized...
> taking a look.... ugh... I have to say I think this is overloading RSS
> beyond what is appropriate.
>
> I thought RSS was for *original* content authors to be able to provide
> notifications about what they are publishing?
>
> If we want a description language for general web pages (collections of
> information and links pointing to/from many sources) then I suggest we
need
> another schema.
>
> To me RSS 0.91 already was a bit too verbose, and these large-scale
changes
> in 0.92 just really complicate matters (as well as being semantically
> muddled (are we calling them "link"'s or "url"'s?)).
>
> (I'm not sure we need a new element for "enclosures", why can't they just
> be another item? (the description would indicate their association...) And
> I think the source really should point to the web site/page where the item
> is offered by the original author, the idea of recognition being to drive
> traffic to the site, most people will be taken for quite a loop by pulling
> up XML source...)
>
> >Well it couldn't be centralized, of course. I was thinking more along the
> >lines of a permanent, generated URI for the site, or a unique ID on a
> >per-feed basis.
>
> Then where does the unique ID for the feed come from if it is not
centralized?
>
> -Robert
>
>
>
>