Re: PROV response to comment on bundles and ids

hi paul,

So we don't allow the nesting of bundles to avoid complications about there
> interpretation.  Thus, in prov-n we introduced the notion of document as a
> container for all statements that are not part if a bundle.
>

thanks. that makes sense.

We can add your example as something that is invalid. One needs to
> redeclare the namespace prefix.
>

i think in general it's useful to provide some invalidation pointers. even
something like bundle of bundle should be pointed to as invalid. i just
reread the bundles section in prov-dm it doesn't explicitly rule out
nesting and in prov-constraints it's somewhat indirect. for container
datatypes, it might a good thing to explicitly state nesting considerations.

cheers,

satra

>