Give up the funk

· Syndication, Weblog Concepts

There is an interesting discussion in the comment thread for Dave Winer’s Why I said Movable Type’s RSS support is ‘funky’ post. The title for that post is accurate insofar as Dave explains why he accused Six Apart of not respecting the RSS spec, but maddeningly evasive when it comes ot explaining precisely what he meant by the original statement.
In other threads on other blogs people have speculated about exactly what Dave meant, but Dave has not come out and detailed the problems he sees. The effect is to spread a meme that there is a problem with MT’s RSS without in any way helping to resolve that problem, whatever it is.
In his post, Dave says he wants “all tools to produce the same RSS, modulo differences that are rooted in real differences between the products.” (Again, without saying how MT fails to meet this goal.)
He states that “not enabling interop is a chicken-shit way to compete. It’s a sure sign of a large installed base but an inadequate development team or codebase.” This smells like a further slam against Six Apart again without any specifics about whether or how this is going on. When pressed, he can easily say he was just stating principles or philosophy, but under the heading of explaining his assessment of MT’s RSS support, the implication of the quoted clause above is clear.
Skipping past the details again, Dave suggests that Six Apart has “responded with some respect, at least now their support for 0.91 has been upgraded to 2.0,” (a situation that preceded this tempest-in-a-teapot and implies that Dave’s agitation resulted in an RSS upgrade for Movable Type, when the trail through other blog comments suggests instead that Dave was basing some of his criticism on the RSS 0.9x feeds produced by default by Movable Type some versions ago – again without clarifying what may have been wrong with that RSS).
Then we get to the meat of the complaint: “…but they still, by default produce RDF where RSS is called for.” What does Dave mean by this? Does he know that MT by default produces both an RDF (RSS 1.0)-style feed and an RSS (2.0)-style feed? Is his complaint that MT offers RDF as the default syndication feed link in the default MT template? If so, is this some sort of unethical bait-and-switch (is that why MT’s default template labels that link “Syndicate this site (XML)” and not “RSS” and doesn’t use the orange RSS icon Bryan Bell designed for Dave? We can have a good discussion of this if Dave will come out and identify this as the problem (or discount it as a side issue and not the real problem, which is?).
Does it have to do with the header information in default MT templates that directs aggregators to the RSS 1.0 feed? Again, without knowing it’s hard to have a grounded discussion about what is kosher and what isn’t.
It is strange to me that Dave writes, “At some point the discussion has to stop, and for me, that was the day I said publicly what I had been saying privately – MT’s support for RSS is funky….” That’s when the conversation stops? Upon attacking an upstart competitor in your blog? Come on.
He goes on “…and guys and gals, I was being kind. I could have said it’s wrong.”
What is wrong?! So far I don’t see it, and I pay more attention to most of this stuff than most people do. If you have to be an ultratechnical insider or privy to private email debates to follow this point (and some pretty smart well informed people all asked Dave the same thing – please be more specific), then what is the point of airing the argument in public without support. It feels like spin.
Dave says that MT’s approach to RSS is “as wrong as it would be for UserLand to implement Trackback that doesn’t work with Movable Type” (so does this mean that the RSS put out by MT blogs doesn’t work with Radio? I have not observed this.)
He goes on: “…or implement the Blogger API and change the order of the parameters.” That’s ironic, given Evan Williams’ unanswered narration of how the MetaWeblog API discarded parameters from the Blogger API crucial for hosted-server blog providers like Pyra.
He end with more kerfluffle, lecturing the other principals as if from a position of greater maturity, saying, “UserLand clearly moved first in RSS in blogging tools, and it’s up to the people who are following, to do so with respect.” (Please, Dave, detail the lack of respect, either the original disrespect or the gap now between some respect and whatever a full measure of respect would mean.)
The funny thing is that I believe a majority of the blog world would agree that the blog software products of today should ionterpoperate smoothly (in most ways, especially when it comes to migration, they most assuredly do not), and that the real threat will come from a Microsoft or other large vendor embracing and extending the emerging weblog standards. The effect of this hullabaloo, however, has been to sow dissension and distrust within the ranks of the still small close-knit blog development community, and somehow (as when he pops into the comments and responds selectively to the questions posted there) Dave seems to be contributing more confusion, vagueness, evasiveness, and big-picture changes of the subject than a simple explanation of what’s wrong with Movable Type’s RSS and what would be required to fix it.

Note: I’ve posted several comments to that comment thread. This blog entry is an attempt to capture some of the issues being discussed there. Here, for the record, are my posts to the thread (slightly edited for context):

Dave, I came to this page eager to understand the nature of the funkiness and came out just as confused.

Is the problem you cite the use of RDF elements in RSS 2.0 feeds? And if that is what you’re talkign about, is it being done in some way that violates the extensibility (as I understand it) of RSS 2.0.

You also mention the metaweblog API. You generously pointed to Evhead’s discussion of it in his blog a while back but you never responded (that I could see) to his contention that the metaweblog API discarded some elements of the spec that Pyra needed for its hosted blogs, and thus “broke” their spec and showed the kind of disrespect and failure to interoperate that you decry here in this post.

Also , it seems like the issue re MT is now not some misformulated RSS 0.91 but the defaulting to RDF (aka the branching from RSS unfortunately known as “RSS 1.0”) vs. RSS 2.0.

Would this remaining “lack of respect” problem be solved by labeling the feed an RDF feed, or would it still gall you as a vote for a competing format? For that matter, is it disrespect for RSS 0.91 or RSS 2.0 to prefer RSS 1.0 (RDF)?

Thanks!
xian 6/14/2003 1:35:28 AM

Dave, you’re responding with philosophy and flummery when no one is disagreeing with you (in this thread) about the beautifies of not locking out competition by “embracing and extending” standards and popular formats maliciously. That’s not in question here.

People are simply asking you to be clear about what is funky about MT’s RSS support. Several candidates have emerged in this thread:

  1. MT default templates offer a syndication link to an RDF-style (aka RSS 1.0) feed.
  2. MT default templates include a header item that offers the RDF-style feed for autodiscovery.
  3. MT RSS 0.9x templates used to include something that violated the RSS 0.9x spec?
  4. MT RSS 2.0 templates still include something that violates the RSS 2.0 spec?

Please simplify clarify whether the problem with MT’s RSS support is one or more of those four issues or indeed something else entirely. If 3 or 4, please indicate the nature of the spec breakage. By being vague, you give the appearance of constantly shifting the justification for the original criticism of Six Apart.

Legitimate specific criticism could do some god by setting a benchmark to help others (including nonexperts) judge if Six Apart is competing fairly in this heavily self-conscious marketplace. Popping the stack when people press you for details does little to reinforce the impact of the ideals you stated.

By comparison, Ev gave specific details when asserting that the metaweblog api “broke” the blogger api.

xian 6/15/2003 3:12:24 PM