Hi, folks-
Sorry for the late comments... I have been on vacation.
At TPAC, I sat in on your F2F sporadically, and commented there on some suggestions to make the API a more attractive implementation prospect for browsers; some subsequent informal discussions with browser vendors seem to bear my intuition out. I've been pretty busy in the intervening months, and haven't had the time to properly follow up with you, for which I apologize; however, I'm really interested in a successful outcome for your work, so I thought it might help to start simple and build out from there.
I recognize that you have done a lot of work on your documents, so I
don't want you to take my critiques the wrong way; I think there is a
lot of good stuff there, which can form the basis of a really compelling set of functionality.
Just to start the ball rolling, here are some stray thoughts I had while skimming your two documents, Ontology for Media Resource 1.0 [1] and API for Media Resource 1.0 [2].
Ontology:
As an editorial comment, there seems to be an academic tone here, with
the use of the word "our" rather than "this specification", detailed
rationales for decisions (which is good in itself, but ), and a
generally tentativeness ("Although the set of properties is now limited,
it already constitutes a proof of concept", section 4.1.1, "proof-read
our interpretation", etc.). I recommend you simply state in the Status
section that feedback is welcome (with short inline notes commenting on
which sections are in particular need of feedback), that there may be
considerations for possible future versions of the spec, and that you
leave room for extensions; if this is done right and sees uptake, it
will almost certainly be the first of a lineage of specs.
1 Introduction
The introduction could benefit by trimming it down. Split the
relationship to Dublin Core into a subsection. Explain the uses of this
ontology to the expected readers of the spec: possible implementers,
content authors, and users of the ontology.
1.1 Purpose of this specification
After reading this, I'm left wondering whether this ontology is expected
to be used in metadata itself, or if it is only a mapping. If someone
were to use this ontology by itself, would that be a misuse? Explain
why or why not in this section.
4.1.2 Core properties
All the property names are prefixed with "ma:", which could be confused
as part of the property name. Simply stating that the properties are in
the Media Annotations namespace is enough (as long as you provide
concrete examples of use).
4.2.1 Rationale regarding the mapping table
"Its namespace is "ma", for Media Annotation." The spec seems to
conflate the namespace with the prefix; usually, a namespace is
something like "http://w3.org/MediaAnnotations/", which is often bound
in a serialized document with a common prefix, like "ma:" using a
namespace declaration; the prefix is not considered universal. (In my
opinion, this is a flawed design for Namespaces in XML, but that's the
convention.)
4.2.2 The mapping table
I really like the level of detail this spec goes into for performing the
mapping (though I guess it's still a work in progress. The mappings
seem a bit hidden, though, and they are really the meat of the spec. I
assume you are trying to keep the spec manageably short, but I would
suggest either keeping the tables inline in the body of the single-page
spec, or splitting it out into chapters with each chapter a short
description of the mapped ontology, followed by the table mapping itself.