peter royal wrote:
> On Nov 7, 2003, at 5:37 PM, Stephen McConnell wrote:
> > Searching is certainly on the my agenda.  We need it for IDE related
> > development, repository management, and intelligent query based
> > artifact aquisition (and HTTP in this context is not an ideal
> > solution).

> But that's a layer that would sit on top of the repository, right?

That is my thought as well.  First we need a layout, of which I like Tim's
the best so far.  On top of that we can add meta-data, and then tools that
use the meta-data.

For "dumb" servers, we can have tools that create static meta-data, which
would also be more efficient.  "Smart" servers might support dynamically
creating, or simply reformatting, meta-data for different clients, e.g.,
Maven, Avalon or JNLP, from a single source, but otherwise we just store
meta-data statically.  The real "smarts" would be on the client side, where
the metadata would be most often used.

But since the meta-data would be present at a URL within the layout, I don't
think that we need to deal with it right now.  Just the layout, itself,
which can be presented to infrastructure and the projects.  Once we've a
layout, existing tools can make use of it, and people can start discussing
the next layer, which would be meta-data.

At least that's my view.  :-)

        --- Noel

Reply via email to