My hubris exposed
by ZetaGecko | Add Your Comments | Atom/RSS
I've posted version 1.0.1 of the Info Bite List specification, exposing my hubris in declaring the previous specification version 1.0. At issue was the fact that the link element had a type attribute which referred to the MIME type of the object being linked to, but, as with any other element which can have a value, the type attribute also referred to the MIME type of the element value. I changed the name of the attribute for referring to the type of the linked object to "mediatype" to resolve the conflict. For consistency, I made the same change for the media element, although, because media elements don't have values, there was no conflict.
I also added more explanation to the section describing the mode attribute based on the research outlined in my two previous posts.
I've discovered two more issues that I may need to revisit, both of which may lead to changes in order to ease (or even enable) the job of validators: valid values for some elements may be different based on the value of an attribute (the rel attribute in a few of the elements), and the number of some elements that are allowed is based on the number of valid values for an attribute (once again, rel). I'll do a little more research and decide whether further changes will be made based on those issues.