intertwingly

Sam, my friend, you can have modules.

On my walk yesterday I thought maybe I should just add them to 0.94.

Then I thought -- I need a use-case. I can't add something unless I can test it.

So if you want to take the lead on this, and provide both the copy to be added to the 0.94 spec, and one or two conforming examples, I'll do my best to get it in. Otherwise let's shoot for 0.95. I want to start writing code for 0.94 asap and next week is very busy. That and only being able to work four or so hours a day are my major constraints.

Hi Dave!

Use case: dc:date

In short, don't add anything more to 0.92. All new additions go into a namespace.

If this approach is acceptable, I'll gladly strip all 0.92 content from 0.94 and invent a namespace for the new stuff.

- Sam Ruby

Sam, I didn't do all this work to undo it. Sorry.

That wasn't the idea. There are people already writing code for 0.94.

Let's look beyond 0.94.

PS: It would be great if your comment system remembered my name, email and url.

Dave,

I'm confused. You put out a draft to solicit input. You explicitly tell people NOT to base any implementations based on what they see just yet:

http://backend.userland.com/rss#thisIsADraft

When is the next window and how do I avoid missing it?

P.S. I'll look into adding *shudder* cookies.


Posted by Sam Ruby

If RSS 0.94/5/6 is going to (a) maintain a simpler core than RSS+RDF, (b) support modules, and (c) stay backward compatible, I would say this is an evolutionary leap above RSS+RDF. All the extensibility without the bloat. You should call it RSS 2.0. (I'm not kidding.)

More on this idea (including the very suggestion of calling it RSS 2.0): http://www.syndic8.com/~wkearney/blogs/syndic8/archives/000017.html

Rael mocked up RSS 2.0 over 2 years ago; he just didn't know what to call it.

http://www.oreillynet.com/~rael/data/xml/rss/modular/demo/


Posted by Mark Pilgrim

Sam, I outlined two choices for the sequence of events on Scripting News this morning. I know the disclaimers are there, and UserLand has not deployed anything -- that's the only developer I have any direct influence over. Others are moving forward in preparation. It always works that way. I've never been able to find a way around it. Basically RSS has no business evolving so much at this late stage in its life. I didn't know Dublin Core is only four years old. RSS is almost one year older.

Mark, thanks for the prespective. I'm a little too close to the situation to know how other people see it.

It's 5:30PM and believe it or not, <i>no one has objected to the plan outlined on Scripting News.</i> In fact, quite a few people have emailed to express their hearty approval. So we're going to go for it. Namespaces in 0.94. And I think that's just the beginning of the coalescing. Just the beginning. Ye-hi.

Hang on there Dave. You need to give some of us a chance.

I'm all for namespaces. (I wholeheartedly concur with Jon Udell's viewpoint in "The mission of RSS.") I disagree with "[elements not defined in the spec] can be in namespaces or not, at your pleasure" quite strongly. The "or not" is a bad idea because its a recipe for namespace collision disaster. The usual examples apply. Its just not worth the risk. I also think there is something to be said for the "cue" a namespace provides indication to what is part of the core and what is an extension.

I want no part of the acrimony that has been leveled at you on this, but I sincerely wish you would post thoughts/plans like this to to an open many to many forum like mailing list. There is a near silent majority that doesn't have the time to monitor dozens of weblogs in realtime for topical information and provide perspective that would be helpful to this effort.

I want to talk about prior art.  But I can't do that yet, because first I need to give my opinion about this "funky RSS" business.... [more]

Trackback from dive into mark

Mark Pilgrim dives into the RSS date flap... "It’s FUD. It’s crap. Cut it out. To recap: in 1982, RFC 822 defined a date format. In 1997, Dave Winer respected prior art by using that date format for the date elements in his syndication...