Re: Review of section 6

On 13 apr 2009, at 23:42, Raphaël Troncy wrote:

> Dear Silvia,
>
> [I suspect Jack has already addressed some of these comments, but I  
> made another run :-)]

Nope, I hadn't: it was still easter here. Thanks for fixing these!

I made a quick pass and fixed one other minor one.

These require discussion, however:

>
>> Section 6.4, Second Editorial Note
>> * we should specify that if named addressing is used, all others  
>> are ignored
>
> I added:
> "Incompatible: if the named dimension is used, all the other  
> dimensions are ignored."
>
>> * we should specify that if multiple temporal or spatial or track
>> parameters are used, only the first one is interpreted
>
> Partially captured by the semantic error #1, slightly rephrased.
>
>> * we should specify that any dimensions specified outside the media
>> resource (e.g. end time after end of file, coordinates outside the
>> frame) are to be clipped to the actual size of the resource
>
> Partially captured by the semantic error #3 in the ed note, but I  
> have completed it.
>
>> * we should specify that unknown id or unknown track specifier are to
>> result in no-op
>
> Captured by the semantic error #2 in the ed note.
>
>> * we should specify that non-sensical time specs or coordinates (e.g.
>> t=20,10) result in no-op
>
> Captured by the semantic error #4.
> Thanks for the carreful reading.

The question boils down to what our preferred error treatment is:  
masking the error and attempting to continue or flagging it.

My preference is that hard syntax errors are flagged (such as  
overspecifying a dimension, or specifying both name and something  
else), and soft errors, which depend on the underlying media item, are  
masked.
--
Jack Jansen, <Jack.Jansen@cwi.nl>, http://www.cwi.nl/~jack
If I can't dance I don't want to be part of your revolution -- Emma  
Goldman