[GEEK] OPML Validator


I just came across Dave Winer's new, beta, OPML Validator, which is a welcome development. While there is an OPML spec, having a validator too is a great, pragmatic, concrete way to give a big thumbs up or down to a particular piece of XML


From the very start BlogBridge  has supported OPML for import and export, as well as for various internal representations and communications. We obviously have done our best to write valid OPML, as best as we understood what that was.


Still, what trumps  every theoretical position on OPML validity are user cries of:  "Hey, I can't import your OPML into XXX", or "Hey, you fail to import my OPML, but YYY has no problem with it." Our approach to this situation has been to follow what has been called "Postel's Law" (which I also learned from Dave Winer, ages ago): "Be liberal in what you accept, and conservative in what you send."


And this has worked well so far, although there have been some sticky wickets that we've had to deal with. You end up feeling like you are chasing your tail trying to accommodate all the odd dialects.


So, it's great that Dave has now put forward the OPML validator. Running our stuff through it, we do quite well. Click here to see the results of running the validator against a pretty big example of our opml.


There are just two errors, each repeated multiple times:


  1. "An <outline> element with more than just a "text" attribute should have a "type" attribute indicating how the other attributes are to be interpreted." 

  2. "An <outline> element should only have known attributes."




In each case we have specified additional attributes  beyond the basic ones. Specifically they are: rating, queryParam, queryType, tags, tagsDescription, tagsExtended and icon.


Some of those probably deserve to become part of the core set, while others somehow will have to become part of an extensibility model. We'll participate in the discussions as they develop. More to come!


Technorati Tags: ,

Posted on October 30, 2005 and filed under Technology.