RSS Is Push?

Remember the famous Wired Magazine feature declaring “Push” the next big thing? (March 1997). Push was going to be so significant it would kill the browser:

“The Web browser itself is about to croak. And good riddance. In its place…”

The feature is famous both for exemplifying Wired’s tendency to make huge, sweeping declarations and for being so painfully wrong (the public ignored push, PointCast died a painful death, and Wired scraped egg from face).

Now, seven years later (Wired 12.05, not online), the magazine has the cojones to run a story “The Return of Push,” asserting that the idea’s time has finally come. Author Gary Wolf makes the case that the rapid rise of RSS has finally proved Wired right. There’s only one problem: RSS is a pull technology. Just sent this letter to

Gary Wolf is right about one thing (“The Return of Push,” Wired 12.05): RSS is fulfilling some of the original promise of push. But that doesn’t mean RSS is an example of a push technology. If I leave a plate of cookies on my doorstep and invite you to come take one every hour, would you then say that I brought you cookies?

In order to call something “push,” the publisher has to willfully send it to the user (and, ipso facto, to know something about that user). Thus, an email newsletter is an example of a push technology. In contrast, RSS “feeds” sit on a plain vanilla web server waiting for an anonymous client to come pick them up. This is plain old http and apache we’re talking about – no magic “push” protocol makes RSS delivery different from the rest of the Web. I request a web page; I pull it toward me. RSS works exactly the same way. If RSS is a push technology, then so is Tim-Berner’s Lee’s original web.

RSS probably is the Net’s next big thing. But it sure isn’t push.

Music: The Cramps :: Muleskinner Blues

10 Replies to “RSS Is Push?”

  1. You know, I’m a bit late jumping on the RSS bandwagon but now that I’ve got my feed reader shoving site updates into my face every 15 minutes, I don’t know how I managed to keep up to date on everything before!

    *cough* Actually, I mostly use my RSS reader for keeping track of new releases on Within 15 minutes of a new episode being posted, I’m on it with my Bittorrent client. Now, if only the feed reader were scriptable, I could have it just watch for certain titles from certain fansubbers and automatically feed the torrent URLs into Azureus. Bam! Instant anime fansub subscription service!

    I need more hard drive space…

  2. I think you may be being too literal in the “push” vs. “pull” argument.

    While the polling archetecture of RSS is not technically a push technology, I would be willing to argue that by “subscribing” (common termonology) to an rss “feed”, one is asking that content be pushed.

    The main difference is that one chooses a subject, and content is chosen for them (even if it is typically filtered by topic). Once one has subscribed, there is no onus to go out and find anything.

  3. Weird.

    I was sitting on the potty and this post materialized on the toilet paper.

    No comments on how I’m compos(t)ing this reply.

  4. Adam, my point is that when you subscribe to an RSS “feed” you are asking your client (e.g. your browser or aggregator) to pull content on a regular basis.

    Adding an egg timer does not convert document delivery from pull to push.

    Visiting a subscribed site in an aggregator is pretty much exactly the same as visiting a bookmarked site with a web browser, both technically and functionally.

  5. Actually, I would leave mneptok alone for a while, I don’t think you want details on THAT situation.

  6. Pingback: unbounded
  7. Great site, Scot. Your blog came up in Google three times in the past couple of days of stuff I’ve been doing. :-) About push, it’s funny how hype = kiss of death for so many technologies. But the good tech keeps evolving, popping up in different forms. ‘member stuff like CASE?

    Again, nice site. Lots of diverse topics!

  8. Pingback: Innovation Matters

Leave a Reply

Your email address will not be published. Required fields are marked *