Chris Ridd wrote: > On 14 Jul 2008, at 16:11, Shawn Walker wrote: > >> 2008/7/14 Chris Ridd <[EMAIL PROTECTED]>: >>> On 14 Jul 2008, at 08:09, Chris Ridd wrote: >>> >>>> On 14 Jul 2008, at 07:46, Shawn Walker wrote: >>>> >>>>> As far as I know, the RSS/Atom format doesn't work like that. See >>>>> section 6.3: >>>>> http://tools.ietf.org/html/rfc4287 >>>> Not even using XML namespaces? I'm sure I've seen feeds containing >>>> Dublin Core metadata using that technique. I'll dig out some >>>> examples >>>> later if you can't find any. >>> All the feeds I've been looking through use RSS, darn it, but my >>> reading of 6.3 of RFC 4287 is that "foreign markup" is explicitly >>> permitted and processors are required not to break when they see it. >> Yes, but if you read section 6.3 of the rfc, it also says that they >> may not change their behaviour if they see it. > > I'm pretty sure that text just means don't break if there are bits you > don't know about. How else would Atom extensions like the ones in RFCs > 4685 and 5005 work?
I know. It's almost like the border-box model problem with the CSS spec all over again :-) The spec seems to indicate that clients shouldn't change their behaviour based on the presence of foreign markup. What that means is open to a variety of interpretations. Regardless, I'm still inclined to think that we want the standard RSS feed strictly to be for human consumption. I'd rather provide an extended, alternate feed for those attempting to automate systems. >> So it seems rather useless to me. >> >> I would rather provide a separate feed for the explicit purpose of >> parsing instead of adding additional overhead to an RSS Feed that most >> readers will completely ignore. >> >> That's functionality that can be added later in the form of >> "/feed/0/extended" or some such thing. > > Fine by me. I haven't got a reasonable use case for adding metadata > formally into the feed anyway. Thanks for the feedback, -- Shawn Walker _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
