OpenIDProblems with RSS
Tim Bray explains RSS to an imaginary bank manager:
Mr Safe: Well, I’ve learned that sometimes we have to take little risks on these Internet standards, and they pay off. RSS is an Internet Standard, right?
Tim: Well, er, actually, no standards organization has actually blessed RSS. But lots of Internet technology comes up from the grassroots like this, you know.
Mr Safe: Yes, you kids have a phrase, don’t you: “Rough consensus and running code,” wasn’t that it? I have to admire the way you engineers work together to get things done.
Tim: Well, actually, there are at least two competing versions of RSS, but we seem to interoperate OK.
Mr Safe: Yes, and I suppose the people behind them are working together constructively to bring order to the landscape.
Tim: Well, actually, the dialogue tends to devolve into childish name-calling. But the stuff does still seem to work really well.
It’s pretty much RSS in a nut-shell :)
Dave Winer - 20th June 2003 16:13 - #
Simon Willison - 20th June 2003 17:29 - #
Please pardon my ignorance -- I realize this is a heated topic, and I don't want to start a flame war -- but can someone please, in one sentence, answer this question in terms that your average blogger would understand:
I've STFW'd to no avail. Every discussion gets bogged down in minutiae and politics. This can't be so complicated that it can't be put in layman's terms.
Joe Grossberg - 20th June 2003 17:37 - #
I dunno, I've implemented RSS twice before. Its not really rocket science. I find most of the compliants I hear are from by people who haven't ever implemented a "hard" protocol. Write a TCP/IP stack then complain about how hard it is to support RSS. Hell, write an HTTP implementation that is cross system compliant. Then complain about RSS.
However, since we're talking about the ideal world here. One thing that would've made things easier for me, as an implementor, is a barebones RSS document. A simple, very stupid sample that represented all that was required to have a valid, standards compliant RSS 2.0 feed (I've chosen not to support RSS 1.0.) Once I had such a feed, the RSS 2.0 specification would be perfect. Just choose the extra elements, and plop them in where necessary. Rinse and Revalidate.
Sterling Hughes - 20th June 2003 17:46 - #
Joe, I think I can try to field that one. RSS is the syndication format - it consits of an XML document with a channel section describing the feed, and a bunch of items decribing each item in the feed. There are numerous different versions but they all have channel information and item details in common.
RDF is something very different. It's a way of representing "triples" of metadata - stuff like [Simon Willison] [is the author of] [http://simon.incutio.com/]. It's the cornerstone of the proposed semantic web, and is somewhat tricky to get your head around (there's a good thread going on Diveintomark at the moment about it).
The reason the two get confused is that RSS 1.0 (which is just one of the several competing RSS versions) attempted to combine RSS with RDF, as it was thought that would make the specification more extensible.
It's all a bit of a mess really. Mark Pilgrim's History of the RSS fork goes a long way towards untangling the difference between RSS 1.0 and RSS 0.9x - RSS 2.0 is Dave Winer's most recent version of the spec which has been gaining a lot of mindshare recently, mainly because it does everything RSS needs to do and can be extended via namespaces, aka funk.
Simon Willison - 20th June 2003 18:41 - #
Joe Grossberg - 20th June 2003 18:47 - #
Daniel Glazman - 20th June 2003 21:40 - #
Two proposals for discussion:
Ken MacLeod - 20th June 2003 23:16 - #
it does everything RSS needs to do and can be extended via namespaces, aka funk.
That's it? That's the "funk" Dave's on about? Namespaces?
scottandrew - 21st June 2003 00:15 - #
Scott: I think so.
Everyone else: this is linked from Scripting News now as "a quiet conversation on RSS" - so be nice :)
Simon Willison - 21st June 2003 00:33 - #
Boy, that selection of "related links" is kind of weird.
Speaking only for myself I'd like there to be a syndication format which is
Agreed that RSS "as she are spoke" isn't that hard to implement. But the politics are really hard, and (as I argued) apt to frighten outsiders.
Tim Bray - 21st June 2003 01:04 - #
Simon, I wanted to try an experiment to see if we could set expectations high.
Since you started out on the high road, I thought let's see if we can stay there.
Tim, what does 100 percent vendor neutral mean?Dave Winer - 21st June 2003 01:17 - #
Dave: Like Caesar's wife, or a Supreme Court justice. Not only pure, but seen to be pure. No suspicion of ownership, control, or undue influence by any interested vendor.
The usual way to achieve this is to get the stamp of a "Standards Body." But that's slow, bureaucratic, and irritating. Is there a better way?
Tim Bray - 21st June 2003 02:30 - #
IMHO, now that the core is frozen, there is less of a concern of vendor neutrality for that portion. At this point, the dialog moves on to the vendor neutrality in the development, documentation, and advocacy of modules.
A clean and thorough specification will likely be a harder nut to crack. This will be difficult to achieve without tightening up the specification.
Sam Ruby - 21st June 2003 02:44 - #
Hehe, they do appear to be a bit biased in this case. They're generated using MySQL's full-text indexing so I don't have any control over what comes up.
Simon Willison - 21st June 2003 08:55 - #
Noel D. Jackson - 22nd June 2003 08:14 - #