Media Annotations API implementation - asking for confirmation and a few questions

Hello all,

I had a look at the Strawman API design and notes Wiki
http://www.w3.org/2008/WebVideo/Annotations/wiki/Strawman_API_design_and_notes
and made up a user interface design based on the
"get-mawg-unstructured-value" method. Could you please have a look at
http://www.sasakiatcf.com/felix/ma/02/
and tell me if I have interpreted the strawman correctly?

Also, I have a few questions:

1) What is a fragment-indicator?

2) Where are the subtype definitions needed for the sub-type-filter ?
I had a look at the mapping table but could not find them.

3) The language-code-filter is a bit unclear to me.
- RFC 4647 http://www.ietf.org/rfc/rfc4646.txt defines two matching
schemes for language tags: lookup or filtering. Lookup returns one or
no value ("give me the resource if it is French"), and is often
combined with priority information (e.g. "French" with priority 1.0,
"English" with priority 0.5, ...). Filtering returns a possibly empty
set of values ("give me all resources which are French"). Is your
intention here really "only" filtering?

- For filtering there are two variants: basic filtering and extended
filtering. Basic filtering matches the language tag against a
so-called basic language range. For a match the language range needs
to be
	identical to the language tag, e.g. the language range "en" = the
language tag "en" or
	a prefix, e.g. the language range "en" = the prefix of the language
tag "en-us".
Extended filtering uses extended language ranges, that is a
specification of an ordered sequence of sub tags. For example, the
extended language range "de-de" matches de-de, de-latn-de, de-latf-de,
de-de-1901, since the sequence "de-de" is always given. However, it
does not match "de-x-de".

My question is what we want: basic filtering or also extended filtering?

Many thanks for the feedback in advance. Note that I will be on
holiday now for two weeks, so my follow up might take some time.

Best,

Felix