http://microformats.org/wiki/microformats-2-parsing
If you have problems with micropub, file issues with the spec.
About 80% of the code I wrote was just dedicated to finding out whether or not I'm updating an article, creating a new one, deleting the article entirely and which attributes need updating.
The JSON format has an immense amount of issues relating to irregularity, some fields have html and text sub-attributes but this doesn't seem documented properly anywhere, not even the MF2 page, everything is an array even where it makes 0 sense to do so and as an implementer I'm somehow expected to then process and store this mess as a coherent article.
Part of the spec validation involves storing a publication which contains almost no text and some user defined elements for some reason. And I'm expected to understand and properly update these custom elements if requested.
My favorite nitpick is that it's rather difficult to tell if a Micropub POST request is an create, read, update or delete operation. They could have used POST, PUT, DELETE HTTP methods but apparently that's too complicated.