1. Jul 30th, 2005

    Are microformats going against the grain

    Robert Sayre responds to Tantek Çelik:

    I disagree. Namespaces can sometimes be a technical hassle, but I’ve noticed that people who rail against them are usually squatting on a set of short strings. Would it be OK for Microsoft to define the meaning of the word “friend?”

    I’m just starting to get involved with microformats. Microformats are content annotations, much easier to use than adjunct schemas. This post is content annotated. The content is all here in HTML for you to read, there are no mysterious fields hidden away from your browser. Part of that content is annotated as mata-data (the tags) for other tools to process. But it’s all the same data.

    That part microformats get right.

    But it also goes against the grain of the Web. The Web is distributed, microformat are not (but microcontent definitely is). Distribution is the opposite of ‘one place to rule them all’.

    Mobile is messaging, fun and games. Not it’s not. Mobile is search and blogging. Not it’s not. Mobile is really all about this.

    Lose the contexts, and you lose the ability of many people to contribute to the conversation by proposing their own formats and extensions. Only a central authority can see value in that.

    Microformats borrow from the W3C (HTML, namespaces) and the IETF (vCalendar, media-types), two organizations responsible for many standards at the core of the distributed Web. That’s distributed by association, not distributed in practice.

    Tags: microformats namespaces lesscode

    Posted by Assaf Filed in general

    1. Jul 31st, 2005

      Eran

      i’m not sure i got your point, how are microformats not distributed?

    2. Jul 31st, 2005

      Assaf

      take the ‘body’ element as an example. it wraps the body of an HTML element. it wraps the body of a SOAP message. it’s used in Open Office XML format, DRML, WDVL, just google it to see how many specs use that one name.

      so how do you get these people to talk to each other and prevent name collisions? you don’t. the fully qualified name for HTML(4.0)’s body element is actually {http://www.w3.org/TR/REC-html40}body. which is a different namespace than SOAP. so each spec writer only cares about shortname collisions in their own namespace.

      microformats don’t have that. so if you create a microformat with the shortname ‘blog’ that contains information about a blog (metadata), and I create a microformat with the shortname ‘blog’ that’s a blog summary (content), then we’re in conflict. and the software can’t tell which is which. there’s a collision.

      collisions are resolved by coordination, coordination requires central authority, so someone is responsible for handing out distinct names, and central authority is the opposite of distributed.

      if XML specs didn’t allow for distribution we wouldn’t have RSS 1.0 and RSS 2.0, because they both use similar-sounding elements (channel, item, etc). but that’s not a problem for feed readers, there’s enough context there to tell them apart (namespace, mediatype), and you can mix both (and throw in atom 0.3 and atom 1.0) in the same page.

    3. Aug 1st, 2005

      Eran

      not quite right. this really should be anwered by Tantek but i’ll give it a try anyway.

      1. microformats follow something like the guidelines for informal XHTML extensions as described here: http://www.xml.com/pub/a/2002/01/16/xhtml-m12n.html (extending XHTML formally is a huge pain)

      2. microformats are disambiguated by including the appropriate XMDP profile. These profiles can be stored anywhere, currently http://gmpg.org is home to a few of those. so you can have different uformats with the same name or with similar elements but still parse properly based on profile and context.

      Eran.

    4. Aug 1st, 2005

      Assaf

      having a page that specifies two profiles in the head@profile attribute (XMDP) still doesn’t resolve shortname collisions. I can only use one ‘friend’ profile at a time.

      and what if I want to syndicate microcontent and I don’t want to choose The Only True ‘friend’ microformat?

      formal extensions are … not easy … but to quote Robert Sayre , “In this case, RSS favors a little technical complexity in return for less bureacracy and more fairness. Maybe morecode is occasionally worth it.”

    5. Aug 5th, 2005

      Labnotes » More on Microformats

      [...] I talked before about microformats not being distributed, the assumption that there exists some authority to define them. Apparently, in more ways than one. [...]

    Your comment, here ⇓

    Or using OpenID