Re: SUO: Re: SUO Ballot with 2 Questions
Adam,
Yes, that's a good way to state it:
> I wonder if the issue is a matter of emphasis. The notion of a
> registry seems sensible to me. I'd like to think that the way we've
> used a web interface to CVS is consistent with this. We've released
> publicly every version of SUMO since its inception, and tried to
> document all the sources in the CVS comments, or in the SUMO versions
> themselves. This can be viewed at
> <http://ontology.teknowledge.com/cgi-bin/cvsweb.cgi/SUO/Merge.txt>
You can think of a registry as something like a CVS repository
with "structured comments" that would enable software to find and
use the different kinds of metadata in a more systematic way.
For example, you could point at a particular axiom or definition
and ask "Find who wrote/modified that axiom when and why in which
version(s) of which source(s)." And the tools could trace through
the chain of metadata pointers to find the original source, all
intervening modifications, and all pertinent documentation.
The purpose of a metadata standard is to enable independently
developed tools to work on common formats. Anybody could develop
the tools -- universities, research groups, or commercial vendors.
> The focus of the standards effort though seems to me to be not the
> tools we use to capture the content, but the content itself. Setting up
> our CVS-based registry took an irrelevant amount of time in the context
> of the larger project. If someone has a better, free, implemented
> system, that would be great.
I agree. As Matthew and I have been trying to say, the focus
is on the content, and the tools are intended to help us
organize, develop, test, maintain, search, and use the content.
John