RSS: I'd Rather Switch Than Fight

Jason Young posed a question on Workbench recently:

I have followed the whole RSS soap opera from even well before Mark Pilgrim was writing snarky posts about Winer numbers. I've actually taken more than one opportunity to call it "Internet Jerry Springer" among my IT colleagues and others that I was evangelizing syndication (and RSS) to -- and that was prior to the 2006 events with the Advisory Board and what I feel is inexplicable behavior from Winer.
The whole hullabaloo has the group of people I work with (a National web effort targeting information coming out of the US land-grants and the cooperative extension system) thinking of solely Atom, and where we say RSS -- we really are referring to Atom. Our web apps are going to encourage Atom, produce Atom, consume Atom, etc. Almost entirely after watching the continued craziness surrounding its use.
I admire and respect your work, your tenacity, your continued evangelistic zeal in what you do here, and elsewhere in the things you write about. So I guess this whole rambly comment is to ask - why are you even sticking with this RSS thing? Why not take that energy and put it behind something with what seems far less "Springer Factor"?
This isn't a troll. It's a serious curiousity on my part.

As we work through the long-unresolved issues in Really Simple Syndication and spark up long-smoldering flamewars, people keep asking me why I don't just switch to Atom. The format's an Internet standard that has a well-established framework for resolving disputes, unlike RSS, and the developer community isn't mired in more old grudges than a Van Halen reunion tour.

I have switched.

I chose Atom for a recent programming project and will continue using it exclusively unless the RSS Advisory Board succeeds in clarifying the RSS specification. I'm no longer evangelizing RSS, because I think the problems with podcasting and HTML create huge hassles for implementers and make interoperability impossible.

Not long ago, I spoke with a technology journalist about syndication, and he predicted that a frozen specification will ultimately render RSS a legacy format. He expects Atom to eclipse RSS completely within five years.

That's not a bad outcome. Atom's a nice piece of work created through an exhaustive public process -- Tim Bray said recently that 17,000 messages were posted on its mailing list during development. I don't know Atom well yet as a programmer, but I've been pleased with the results so far.

Though I no longer pimp RSS, I'm still serving on the RSS Advisory Board. A recent post to the group's mailing list demonstrates why. Sean Lyndersay, a member of Microsoft's Internet Explorer team, needs to know the preferred MIME type for RSS documents. He's looking for a recommendation so they can follow the advice and check an item off the team's to-do list.

Like many subjects involving RSS, this is a long-standing issue that lacks resolution because there isn't a clear authority over the format. Mark Nottingham made an effort in 2001 to register an official RSS MIME type, but it failed:

application/rss+xml isn't registered, because the IESG wanted a "stable reference" for the spec (it being in the standards tree). So, it's technically incorrect to use it now; this is one of the reasons this is still a confusing issue.

There needs to be a place where the organizations and individuals with the most invested in RSS, both in financial and personal terms, can come together to resolve matters like this. The board works in the open, favors no vendor and is eager to bring in outside viewpoints. If that remains true, it should eventually gain the trust of the RSS community.

I think it's worthwhile to fight for that, even though I'm no longer interested in fighting for RSS.

Rss · Atom · 2006/03/31 · 33 COMMENTS · Link

Comments

This would be more credible if (a) it weren't already April 1st in Shanghai, and (b) if Workbench's feed were Atom 1.0.

#1 · Sam Ruby · 2006/03/31

Many thanks for your continued efforts to wedge some sanity into the discussion, Rogers.

I couldn't help but verify that your "SUBSCRIBE" button still generates RSS 2.0.

#2 · Sterling Camden · 2006/03/31

OK, so it looks like you took your template from last May and put it online. That feed wasn't there a few hours ago.

For the next step, the Feed Validator will be more than happy to give you line by line instructions on how to upgrade to 1.0.

#3 · Sam Ruby · 2006/03/31

Thanks. My weblog's Atom 1.0 feed is now online. As an experiment, I'm routing all RSS requests to this feed. I'm curious about whether aggregators can handle that. The RSS 2.0 code's still around so I can offer both, but I'd prefer to offer a feed in one format so I avoid the need to debug two.

#4 · Rogers Cadenhead · 2006/03/31

My weblog's Atom 1.0 feed is now online.

COOL!

The one warning is due to an missing > on an img tag - the one for dallas-kennedy-memorial.jpg.

#5 · Sam Ruby · 2006/03/31

I can say categorically that I *always* choose Atom over RSS 2.0 when I subscribe to feeds. This is as a direct result of Dave Winer.

#6 · _r · 2006/03/31

underbar R,

The fact the you subscribe to feeds at all is a direct result of Dave Winer.

#7 · Andrew Lasey · 2006/03/31

Andrew Lasey: funny, I could've sworn that RSS was invented at Netscape. RSS became popular *despite* Winer, not because of him.

#8 · _S · 2006/03/31

I've known Winer's a freak since somewhere around 1999.

I really haven't been keeping up with the RSS debacle but I'm surprised some people are just now figuring that out.

Scary.

#9 · _M · 2006/04/01

Your revisionist history is bunk. Are you Mark Pilgirm the virus writer?

#10 · Andrew Lasey · 2006/04/01

This would be more credible if (a) it weren't already April 1st in Shanghai, ...

I overlooked this comment. I'm with Anil Dash on April Fool's Day jokes.

#11 · Rogers Cadenhead · 2006/04/01

April Fool's Day is f-ing stupid. Is this legit, Rogers, or is this an overtly subtle April Fool's joke?

#12 · Brian Carnell · 2006/04/01

It's legit. I'm going forward with Atom over RSS on any project where the choice is up to me. I can't build new software on RSS 2.0 while the spec's arbitrarily frozen to much-needed clarifications.

I wasn't planning to recode any existing feeds, but since Workbench is the place that I experiment, I went ahead and switched to Atom here to see how it affects people who read this site via syndication.

#13 · Rogers Cadenhead · 2006/04/01

Now there you go and break your feed.

You can convert the error into a warning by adding an empty <title/> element.

Spec § 4.1.2 and § 4.1.1.1

Oh, and while the word Subscribe links to your Atom feed, the icon that preceeds it still links to your (defunct) RSS feed.

#14 · Sam Ruby · 2006/04/01

Much better! Thanks!

#15 · Sam Ruby · 2006/04/01

I didn't know title was mandatory. Guess I should read the spec.

#16 · Rogers Cadenhead · 2006/04/01

hmm, your atom feed is breaking in Akregator on KDE... Don't know if it's your fault or Akregator's. Your feed is validating, but I took a look and you're using CDATA escaped html instead of just type=xhtml in the 'content' section

I can't find mention of CDATA in the Atom1.0 spec, but I'm sure someone here knows better than I do...

-p

#17 · PeterMHoward · 2006/04/01

sorry to double post, but just realised why Akregator breaks: the html inside the CDATA tags isn't valid -- there's no opening 'p' tag... Akregator heads for the first opening tag it finds and displays from there (whether it's an img, a, i, etc)

-p

#18 · PeterMHoward · 2006/04/01

Rogers, can you please let us know how this switch goes, from a dev/tech perspective? I assume Atom solves the ambiguous "number of allowed enclosures" and "where is html allowed" issues.

Are there other things you've seen that are worth mentioning for people who have never used Atom and are somewhat used to RSS, other than manditory title ;) ?

#19 · Shannon J Hager · 2006/04/02

Peter:

CDATA is a part of the XML specification, and was expressly designed for the purpose that Rogers is using it.

As to Rogers use of type="html", again Rogers' use is exactly as it is intended by the Atom spec. Specifically, the content of this element is expected to be escaped text and markup which could be validly placed inside an HTML div element. See the example in the Atom specification.

#20 · Sam Ruby · 2006/04/02

After the recent unpleasantness involving legal threats to Rogers, I decided to offer ATOM and Feedburner feeds only. I use WordPress, and asked my ISP to help me. I am not sure it's working correctly, and would appreciate feedback, as I'm afraid that I broke any subscription people had to my RSS feeds. Thx in advance.

#21 · Ron K Jeffries · 2006/04/02

I'll definitely post on Workbench as I run into issues with the switch. I don't know Atom that well, beyond a few skims of the spec.

I chose CDATA to encode HTML in my feed because I don't like the code I've written on Workbench to escape HTML markup (I'm not composing in XHTML at this time). Putting HTML inside CDATA seemed like the best choice until I rewrite that code.

I'll add a starting P tag to entries. Let me know if it resolves the problem in Akregator.

#22 · Rogers Cadenhead · 2006/04/02

It's not so easy a support one or the other argument.

We continue to support both formats on most our blogs. I know this is confusing for some end users who don't understand what to use, but it's the dilemma that has been thrust on every publisher unless they send it through a third party like Feedburner.

So much for 'simple' syndication on any real world user level.

#23 · TDavid · 2006/04/02

i moved to atom1.0 on my wordpress blog,no point in confusing people with rss and atom one feed type is enough

#24 · jck · 2006/04/02

Rogers, sounds like you found a way out from between that rock and hard place ;-)

Any info on your experiences in the transition much appreciated.

TDavid, "It's not so easy a support one or the other argument.". Why not? There are good technical reasons for favouring Atom over RSS 2.0. Are there a significant number of readers/aggregators that don't (or won't in the near future) understand Atom? Is there any other reason for supporting 2 different formats?

#25 · Danny · 2006/04/02

RiverOfNews doesn't support Atom 1.0, but that's just damage to be routed around.

#26 · d.w. · 2006/04/02

Please to help me to understand. You have a website with no enclosure, not podcaster, and you know how CDATA work, you have not read the ATOM spec, and as far as anyone know your RSS is work great for everyone who use it. So you are changing why? Am I only one who things YOU are the one who flaming?? That you are the Jerrer Springy everyon won is talk about? Excuse my poor English, I am Portugese, but even I can smell a rat and you stink Senor Rogers.

#27 · Jorque Rene Torqanada · 2006/04/02

Sam, thanks for your comments re CDATA... I used it back when I did an RSS2 feed, having struggled to find any clear spec and just copied feeds that seemed to work. Never realised it was part of XML. ...I switched my feed entirely to Atom1 a couple months ago (as soon as Akregator supported it!)

And Rogers, I expect an opening tag is all it'll take to get Akregator displaying properly -- will let you and them know if there's a problem. Thanks.

-p

#28 · PeterMHoward · 2006/04/02

Jorque,

I am not Roger, but my interpretation of what he's saying is this: RSS is stuck in the mud because the standard is frozen so eventually people are going to have to move to Atom. Roger is a consultant so it is in his interest to know where things are going and be experienced in that technology.

So in other words, his decision (if I'm right) is based on the fact that he now feels things will move to Atom eventually and he wants to be one of the first there.

That said, Atom does add some inherent advantages just in being a well defined standard and he certainly doesn't lose anything by switching.

#29 · Tom · 2006/04/02

Tom,

Atom is just as frozen as RSS.

Correct if wrong.

#30 · Bien Ben Hue · 2006/04/02

Bien Ben - You actually are wrong. Atom is an open standard which means that revised versions can go through the standards approval process just as the 1.0 standard did. That can't be said for RSS 2.0

#31 · Tom · 2006/04/03

Danny Ayers - you answered your own question with the answer (in the parenthesis):

"Are there a significant number of readers/aggregators that don't (or won't in the near future) understand Atom?"

When the future arrives, just holler.

#32 · TDavid · 2006/04/08

Hey, now you have me worried. We are outputting our Poker articles in RSS format, version 2.0. The links are on our front page Poker Magazine On the other hand, if Atom becomes so popular, I would like to make that available as well. Does anyone know of a RSS to Atom convertor?
Also, can Atom handle full feeds and podcasts? PHP preferred, but Perl would not be so bad. Also, are there any Atom based tools out there?

#33 · Mike · 2006/05/14