Re: RFC: On date and time

["Juan A. Suarez Romero" <jasuarez igalia com>]

On Thu, 2011-06-02 at 14:43 +0200, Guillaume Emont wrote:
> Two metadata keys hold dates:
>  - GRL_METADATA_KEY_DATE, "Publishing or recording date"
>  - GRL_METADATA_KEY_CREATION_DATE, "Creation date"

Now that you mention this question, maybe we should re-consider the
GRL_METADATA_KEY_DATE. Currently, it is an ambiguous key: while
CREATION_DATE has an explicit meaning, the date where the element was
created, it's not the same in the case of DATE. Of course, documentation
says "publishing or recording date", but this meaning is not implicit in
the key name. Moreover, publishing date can be very different than
recording date (and to be honest, recording date resembles quite a lot
to creation date).

My suggestion is to drop the GRL_METADATA_KEY_DATE and replace it by
other less ambiguous keys, like GRL_METADATA_KEY_PUBLISHED_DATE (not
sure if we need this key right now) or
GRL_METADATA_KEY_MODIFICATION_DATE. 


	J.A.

PS: I agree with Guillaume's proposal about using GDateTime instead of
strings. And I think the same should be applied to other keys, like
"last-played-time". And following this same reasoning, I would like to
see "source" key as type of GrlSource, instead of type string.