Centralized subscription? Not that way, thanks.

In a surprisingly ahistorical post, Dave Winer suggests a centralized RSS subscription list service. That seems familiar, though not with exactly the same players, and not with that solution.

Flash back to October 2002: Userland’s Radio was spreading coffee cups linked to, AmphetaDesk was spreading pills linked to, new RSS aggregators were popping up all the time, either piggybacking on a previous aggregator’s chosen port or selecting their own, or thinking that they might like to invent a pseudo-protocol like newsmonster-subscribe: or feed:.

The problem is this:

  1. A feed producer knows the URL for her feed.
  2. A reader has a URL that will add that feed to their aggregator’s list of subscriptions.

Dave’s solution is to have the feed producer pass that URL to his centralized service, and to have the reader’s aggregator give the list of subscriptions to his centralized service, which will combine the two, and make the resulting list available to the reader’s aggregator, and to world+dog, too. If I find a feed I’m interested in at Spastic Colon World (I’m not, but if I was it’s none of your business), I would click a link to tell Dave “Hey! I’m reading about spastic colon problems!” and he would tell my aggregator, and you, and that charming woman I met this morning who is busy googling me right now.

But there’s another way to deal with the space in that ellipsis for number two: if you see the problem as being two small problems, that producers shouldn’t have to individually keep track of all the aggregators and have a link to each, and that readers should have a way to automatically convert a single standard link into whatever form their aggregator requires, then the middleman doesn’t need to do nearly as much, or keep as much data centralized. All the producer needs to do is link to the service with her feed URL as a query parameter, and all the reader needs to do is either choose the correctly reformulated link for his aggregator from a list, or if he doesn’t mind telling the site one small and anonymous fact, which aggregator he uses, then it can set a cookie which will then be sent along with the producer’s URL, and the middleman site can just silently redirect to the aggregator.

The reason I mentioned October 2002 is that I’ve been using that very system since then, thanks to Morten Frederiksen’s Syndication Subscription Service. Click my RSS icon, and you go to a page where I see that completely without my knowing about it, I now have a subscription link for IzyNews along with quite a few others. Let Morten set a cookie, like I just did since mine had disappeared somehow, and now it looks to me like my RSS icon links directly to Bloglines’ subscription page. To you, it might look like it links to Radio, or MyYahoo!, or anything else that can subscribe via a URL.

Morten’s page supports autodiscovery, so if you just copy-paste the link from my icon into most aggregators, they’ll still subscribe just fine, which also makes it a workable solution for Firefox’s Live Bookmarks for sites where the content author doesn’t control what goes in the <head> and thus can’t add an autodiscovery <link>.

Scaling? The link goes to a purl.org redirector, so passing the service off to another server ought to take a half-hour or so from a standstill. I hadn’t actually checked it for a while, so it could have been transferred twice, and be running off a half dozen load-balanced servers since the last time I looked.

I can see a few possible advantages to having a third-party subscription list service, all far outweighed by the privacy and reliability problems, but it just isn’t needed to turn philringnalda.com/index.xml into add.my.yahoo.com/rss?url=http://philringnalda.com/index.xml


Comment by Tim #
2005-01-12 01:30:24

Phil’s Back! Yeah!

Comment by Tim #
2005-01-12 01:51:14

I guess you got BB’s cake.

Comment by Jonathon Delacour #
2005-01-12 05:04:38

Woo hoo! That was a happy moment — to see ”phil ringnalda (1)” in my aggregator.

Welcome back!

Comment by Simon Jessey #
2005-01-12 08:19:05

Good heavens! I thought you’d fallen off the face of the Earth. I hope this a sign that you’ll be posting more often.

Comment by Shelley #
2005-01-12 08:30:53

Hey, I’ll take credit for the cake and the digital file.

Mighty good to see your writing, Phil. We need your commonsense, not to mention your humor in tech. Heck, in life.

Comment by Roger Benningfield #
2005-01-12 10:01:04

One more voice added to the choir… glad you’re back, Phil. Even if it’s only for a post here and there.

And I basically agree with you. Dave’s plan is just more complicated than necessary, at least when it comes to eliminating those ”subscribe to MyYahoo” buttons.

Comment by pb #
2005-01-12 11:17:05

Can someone explain what the problem is? I produce feeds and consume feeds and things seem to work OK today. What problem am I having and how is a centralized service going to help?

Comment by Adam Kalsey #
2005-01-12 14:10:30

Two things slowing the understanding and adoption of RSS by mainstream consumers are that feeds are rendered as raw XML by the browser, meaning that someone clicking on a feed link gets a lot of code they see as gibberish; and that subscribing to a feed is usually a multi-step process of finding a feed, copying the link, opening up the subscription mechanism of the feed reader, and pasting the link in.

How will a centralized service help? Not a bit. We don’t use a centralized service to figure out which application to read HTML pages in, we don’t need one for feeds either.

Existing standards already have a solution for this: http://kalsey.com/2005/01/handling_rss_in_the_browser/

Why don’t we use them?

Comment by Jim Kloss #
2005-01-12 14:03:25

I was so happy to see a new entry that I read the whole thing and really tried to understand it. It feels wonderful to know you’re still there and I’m still unable to follow…

2005-01-12 18:29:44

Like folks was saying, nice to see you back in action, Phil.

Comment by Randy Charles Morin #
2005-01-13 20:12:26

Phil, autodiscovery is great. But we need something where the user isn’t using a browser that understand autodisco.


Comment by tph #
2005-01-13 21:34:52

Like everyone else was saying, nice to hear from ya.

(Not that I’m exactly Mr. Talkative himself…)


Comment by Randy Charles Morin #
2005-01-19 06:36:23

Phil, In response to your concern and other’s concern that USM (mime-type click and subscribe) won’t work, I’ve put together a sample C# application that works. And posted the code.


Hope you like.

Comment by Danny #
2005-01-19 13:04:27

Phil, you’re back just in time! The sky is falling! Circle the wagons! (again)

Trackback by Jäger #
2005-01-12 08:43:12

What Yahoo Problem?

Dave Winer wrote an entry yesterday titled ”The solution to the Yahoo problem” Yahoo sends emails to bloggers with RSS feeds saying, hey if you put this icon on your weblog you’ll get more subscribers. It’s true you will….

Trackback by Adam Kalsey #
2005-01-12 14:11:49

Handling RSS in the browser

People often forget that many of the problems faced by RSS and Atom are not new. They’ve already been solved, so instead of reinventing the wheel we should use the existing standards. How a browser handles an RSS feed is the latest example of this.

Trackback by Jäger #
2005-01-13 14:42:59

The Yahoo Problem (II)

Leslie ”0xDECAFBAD” Orchard writes: But, in the comment above, Phil [ringnalda] makes a suggestion that seems ideal to me. Don’t link to feeds directly, don’t use a funky protocol, link to a ”playlist” of feeds. URLs linking to MP3…

Name (required)
E-mail (required - never shown publicly)
Your Comment (smaller size | larger size)
You may use <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <del datetime="" cite=""> <dd> <dl> <dt> <em> <i> <ins datetime="" cite=""> <kbd> <li> <ol> <p> <pre> <q cite=""> <samp> <strong> <sub> <sup> <ul> in your comment.