> From: Adam R. B. Jack [mailto:[EMAIL PROTECTED] > > > Is there any consensus out there that the > > repository URI proposals are the right/wrong way to go? >
[snip]
>
> I would like to change the name of product-specified (back?) to
> 'group-specifier'. I think that concepts like 'sub-project' are totally
> orthogonal to a artefacts repository, and how folks structure things like
> that is not something we ought attempt to replicate. There are other
> groupings we aren't allowing in, so why allow this one? I think
> that 'group'
> becomes a namespace, and that products within that group (e.g.
> sub-projects)
> can't uniquely name their artefacts with the group. That is all we need.
> Anything else is distraction/overkill.
>
> As such, I'd say group-specifier=pchar* (i.e. no '/' in there, so it is
> parsable.)
>
OK, there are now three alternatives:
1. repository-uri = access-specifier "/" product-specifier "/"
version-specifier "/" artifact-specifier
product-specifier = organisation "/" project
organisation = pchar+
project = pchar+
2. repository-uri = access-specifier "/" product-specifier "/"
version-specifier "/" artifact-specifier
product-specifier = organisation "/" project
product-specifier = path_segments
3. repository-uri = access-specifier "/" group-specifier "/"
version-specifier "/" artifact-specifier
group-specifier = pchar+
The third form leads to a flat repository structure, similar to
that in use by maven (http://www.ibiblio.org/maven)
>From a browsing perspective, this doesn't scale to large numbers
of groups (aka products).
Example attached.
Regards,
Tim
repoform3.tar.gz
Description: Binary data
