iBLOGthere4iM
<< Previous Main Next >>

Everybody keeps telling me that they would be in favor of using RSS 1.0 or RSS 2.0 as the object/data model for Atom API, but that they don't think it's feasible, as the community is anti-RDF. Well, I'm anti-RDF, but I'd prefer RDF to the current Atom syntax. Why re-invent the wheel? Why not respect prior art?

Yes, we all laugh when we hear the term "Semantic Web", but that doesn't mean RSS 1.0 is bad. Sure it has semantics that we don't think will ever take off, but why not give the Semantic Web a chance, regardless of whether you know or not that it's doomed. Give RDF a chance. To think, iM the one that is calling for re-use of RDF. Give me a break!

Why not use RDF? Cause RDF doesn't have constructs that are in the Atom syntax? Then write an extension module and get on w/ writing application instead of re-inventing the wheel. More spec. More spec. More spec. Three syndication formats. Three blogging APIs. Three Web RPCs.

The solution is to merge RSS 1.0 w/ the Atom initiative. Give the whole pie to the IETF and let them (the working group) figure it out.

Update: I think today I finally realized the problem w/ RDF. Talking via several forums w/ RDF people, they don't seem very convinced in RDF. Although they tout, they shy away quickly. There's no real passion in what they are doing. They seem to have a doubt that the RDF thing can work. That's too bad. Dave Winer has passion. I guess that's why he won.

Reader Comments RSSRmail

Thanks Randy for the kind words.

I still think the rational approach for Atom is to start with RSS, and make the changes that must be made to achieve the new functionality they want.

And to do the same with the Blogger+Metaweblog API.

There's a tendency among programmers to want to hit the reset switch just becuase (they think) they can. With Google's support they have the chance to really disrupt things. Movable Type users have much more choice, but Typepad has the same power as Blogger to disrupt.

But the irony is that leaders have the most to lose from such disrpution. Much better for them to grow the market rather than disrupt it.

The questions of elegance are arguable, and imho irrelevant. What matters is what works for users. As we've found out in the focus group on the BloggerCon site, the users are starving for fixes and features. I didn't see one user say the format and editing APIs need to be redone. I think if they understood what Google and MT are doing they'd be very unhappy.

Randy:  I am by no means an RDF expert, but what I've gleaned from reading the blogs of people who are is  that, even among people who are fans of RDF and who are SemWeb proponents, there is general acknowledgement that the RDF-XML serialization (as used in RSS 1.0 and other places) is particularly ugly, unreadable, and tough to generate correctly with the common XML toolkits people are using.

I know that it was an Atom design goal to produce, at least in the serialization, something that was human parseable (indeed, even as a non-coder I was able to produce a validating feed simply by doing a view-source on some other valid feeds and reading the preliminary spec pages), and I think they hit that sweet spot.

<a href="http://www.freeke.org/ffg">Dave Walker</a>

Dave Walker,
Then change RDF Site Summary. The confusing part is the Seq collection. Remove it. It's unnecessary and from what I've heard from RDFers, they agree.

Randy

P.S. Sorry I switch to WYSIWYG on Moz. No more raw HTML entry.

Dave Winer,
I can't believe u link to me calling for use of RDF Site Summary :)

Randy

My impression of the pro-RDF crowd is that they're pretty passionate. They just have trouble dumbing down the concept enough to sell those of us who haven't drunk the Kool-Aid yet. I've tried several times to grok RDF and the Semantic Web, but I always end up fleeing in terror.

Ignore the RDFism behing RDF Site Summary. Focus on the syntax of RDF. It's not that horrifying. Maybe a little ;)

Randy

Randy, the bigger point you were making was that discontinuities suck. I wanted to say something about it, and give people on my blog a chance to comment, so I linked to it.

I also thought your ideas deserve attention because the current approach to Atom is just one of several that should be considered. It's as if a decision has been made by consensus. Nothing could be further from the truth. I wonder how we're going to communicate that to the people who evaluate Sam's proposal to the IETF. Sam sure ain't gonna tell them that there are quite a few approaches that were shouted down by a handful of people who camp out on the mail list and the Wiki. I can't afford the time or money to fly to Korea to tell them personally.

Dave (Winer)-- the IETF WG (whoever ends up on it) will have both the mailing list archives (which are public) and the edit history of the wiki pages (also public) to look at, if they want.  I really hate the way you keep personalizing this as some nefarious plot by Sam, who I've never seen say a bad word about anyone, including you.  Let's keep things technical and civil.

Hi Randy,

It's a shame that there has been all the historic confusion surrounding RSS, but there isn't any chance of unification around a single version - there was a good opportunity with RSS 2.0, but a certain attitude won the day.

So we have to move on, and look to the positive. I think there's a lot more to be gained from analysing the requirements of blogging and syndication in general and coming up with a spec that fits them, rather than emphasizing backwards-compatibility. Discontinuities don't suck as much as maintainence of outdated (or simply bad) design. 

The way the Atom API has emerged is a case in point - it's considerably more straightforward than the APIs that went before. More straightforward means better tools for the end user.

Regarding RSS 1.0, the RDF model is designed for interop, so compatibility with Atom is pretty well guaranteed as long as Atom is tightly spec'ed. Ok, Atom won't be using the RDF/XML syntax, but that isn't such a big deal - it'll still be possible to use Atom data in RDF-based systems.

Personally I would have preferred Atom used RDF/XML directly - I think in the long run it would make things easier - but making the web a better place is the common purpose.

btw, I started to compile some Atom + SemWeb notes, suggestions welcome:

http://dannyayers.com/2003/musings/echo-semweb.htm

One thing just occurred to me that's missing - the difference between Semantic Web technologies (RDF, OWL etc) and the Semantic Web vision. RDF is bloody useful whether or not you subscribe to the vision...

  

I can see the next Gibson biopix now:

"The Passion of The Dave Winer"

-m

No, iM not saying that Atom is a "nefarious plot by Sam." iM saying that this road is one I wouldn't travel. I breaks interop. It's wrong.

"...Sam, I've never seen say a bad word about anyone, including you."

Big time agree w/ "RDF is bloody useful whether or not you subscribe to the vision..." Beyond RDF Site Summary, FOAF is also great. I won't lie and say that I like RSS 1.0 over RSS 2.0, but I'd rather RSS 1.0 over Atom syntax.

Randy

When the RDF/Atom  topic came up last summer, several people took a look at ways to do it, and even worked up a couple of examples.   I believe the end result was a syntax that, while not RDF itself, was regular, normalizable and  transformable into RDF with widely available tools, and that the effort made in this direction was appreciated by the SemWeb proponents.

Randy: looking at what you cited, I don't see an ad hominems or personal attacks, though I appreciate,  being on the receiving end of the posts,  you might have a different POV.

I like the fact that we're discussing syndication and no one's broken out the ICBMs yet in this thread. :)

Just to balance the thread :) CDF (1 ICBM) Sucks!

Randy

Speaking of ICBMs - the US Department of Defense doesn't appear to get the joke about the Semantic Web:

http://www.gcn.com/23_3/interview/24814-1.html

 

I think the lack of ICBMs is from the panzy blue borders :)

Randy

Re: Sam
Just a note that Sam n I work together on the Atom WSDL. I don't want anybody to get the idea that we're enemies because we disagree on the Atom approach.

Randy

Do we have to go to Korea to be part of Atom?

Seems like a pretty bad idea.

Was that voted on somewhere?

 

It'll be voted-on in Korea :(

No you don't have to go to Korea to participate in the development of Atom. You can simply put your thoughts on the Wiki or the mailing list.

Randy

Randy, putting your ideas on the Wiki gets them erased, and putting them on the mail list gets you flamed. I think you're wrong, if you want to participate you have to go to Korea. It wouldn't be so bad if the advocates didn't always say it Atom is an open process. I don't think the IETF would be very happy to find out that's it hasn't been open at all. Basically the IETF is going to rubber stamp an act of authorship as if it were an open standard?

Interesting thoughts. I have to agree w/ the mailing list comment :) By the way, IETF isn't stupid. If enough opinion is voiced, then they will listen.

Randy

>>If enough opinion is voiced

Good. The key thing is to gather the opinion in a way that no one can say they weren't heard, or their ideas were erased, or shouted down.

Basically an alternate proposal should be made to the IETF that says there are other approaches for an IETF-endorsed format and protocol for weblogs that have significant support. I don't think they'll be swayed by the size of Google, or the niceness of the personalities of the proponents of the Ruby-Pilgrim approach. (And we all know that some of them aren't so nice.)

The shouting-down approach to consenus-building is not something that IETF should support, imho.

Every page on the wiki has a clearly annotated edit history, folks.  That's not conjecture, it's fact. 

As far as flaming on mailing lists, well duh, that happens on every mailing list, and is hardly endemic or specific to the Atom process.

For a bunch of folks who claim to be about consensus building and respect for prior art, it's disheartening to see so much stop energy expended here.

The shouting-down approach to consenus-building is not something that IETF should support, imho.

By the same token, people who only poke their heads out on the ML or wiki to snipe at one or another aspect of the spec, without posing alternate solutions or providing the all-important alternative spec text at an earlier point in the process, are going to be ignored.

(http://intertwingly.net/slides/2004/etcon/62.html)

I think there needs to be an effort to guard against certification by the IETF of Atom. As much as the Atom group wishes it, the direction is dictated by few. For instance, iM on vacation and it looks like SOAP support is being removed. Convenient timing.

Randy

that's right randy - it's all about you.

"iM on vacation and it looks like SOAP support is being removed. Convenient timing."

If you're implying that people are planning these decisions around your vacation, and that you alone could have stopped them had you been availble, then you are even more arrogant than most of us would have ever guessed.

that's another nap for lm0 and looks like b needs one too :)

Randy

Hate to tell you, but niether was from "lm0" or "b", whoever they are.