An author writes to ask if its safe to ignore Atom in his weblog, and I say…

An author writes to ask if its safe to ignore Atom in his weblog, and I say absolutely yes it is safe. Look at it this way. Scripting News is the most subscribed to feed in the world. And I promise it will always be available in RSS as it is today, so as long as people want to read my site, the aggregators will have to support RSS 2.0. I can offer the same kind of safety that Lotus 1-2-3 offered developers on MS-DOS or Excel on Mac OS. If you did something the way they did, you were safe, because you could be sure the platform vendor would never break them. In this case what matters is if aggregators read the format. The day aggregators can’t read Scripting News is the day your RSS feed will stop working. My job is to be sure that day never comes.
[Scripting News]

Is this:
(a) A reasonable assumption
(b) King Canute sitting on the shoreline, or
(c) Somewhat egotistical rambling?

  • http://cheerleader.yoz.com/ Yoz

    IMHO, all three:

    (a) Yes, assuming he’s talking about continually outputting exactly this version of RSS 2.0. It’s both way too popular and too easy to support (in a basic way, anyway – see below) to be fully phased out – at least, while we are still using HTTP and HTML as well.

    Plus, he originally said this in June of last year, and that was when he wasn’t as violently anti-Atom as he is now: http://backend.userland.com/2003/06/26#a313

    (b) Yes, because Atom will become the dominant standard within a year or so. Blogger is already outputting it by default, MT and LJ will be soon, as will a billion smaller engines.

    The tide is not just these new outputted feeds – it’s also a whole load of developers who would rather have an unambiguous spec than a simple but vague one. (The ones who prefer a simple but vague one aren’t the ones writing the good libraries)

    (c) Yes, though not for the reasons you’d think. Most of what he’s saying are the things that *should* be said by a responsible developer of standards. The trouble is that the RSS 2.0 is not only a vaguely-specified mess but, as Mark Pilgrim recently showed, actually incompatible with itself:

    http://diveintomark.org/archives/2004/02/04/incompatible-rss

    It is so incompatible with itself that the same tags with the same data can mean two different things. Not only that, but one of the first things in all versions of the RSS 2.0 spec is that an RSS 2.0 file should conform to the XML 1.0 spec, yet scripting.com’s own rss.xml is not valid XML. (Of course, “conform” could mean that it should just be “well-formed” XML rather than “valid” XML, but unfortunately guessing what it means is left as an exercise for the reader. Or implementor.)

    So what Dave calls RSS 2.0 and what your aggregator thinks is RSS 2.0 and what your personal feed producer thinks is RSS 2.0 may all be different things, and if that hasn’t already been the case then it may well be the case at some point in the future if Dave decides, again, to change the spec while keeping the version number the same.

    This problem is further exacerbated by the CC license that the RSS 2.0 spec was recently put under. It allows anyone, not just Dave, to change the spec any way they like and publish it and still call it RSS 2.0. Allowing anyone to change a standard however they like is one of the *worst* things you can do to deflect criticisms of your standard not being open. Sure, it’s now WAY open. It’s just not a standard any more.

    (Bloody hell. I know far, far too much about this. But I just can’t tear my eyes away.)