 |
 |
On Fri, Jan 24, 2014 at 11:16 AM, Robert Sanderson <[log in to unmask]>wrote: > > All in my opinion, and all debatable. I hope that your choice goes well > > for > > > you, > > > > I'd like to repeat: just because I agree with that choice, and I'm > > defending it here, it wasn't my choice. Not at all. And the concerns you > > express were well-aired and very carefully considered before the choice > was > > made. > > > > And yours :) > > Ok, that makes me feel a bit personally defensive... I just want to be sure that it's clear that while I agree with my client, as a developer I'm not happy with opaque URIs for predicates any more than you are. The defaults that I've written into the Open Metadata Registry for coining URIs are: opaque numeric for value vocabularies, and camel-casing of the label in the default language of the vocabulary for predicate vocabularies. I think that's the way it should usually work -- my personal best practice. But these are decisively multilingual vocabularies, without a 'default' language for the labels. It's a French and Spanish and English and Hebrew and Arabic and Italian (etc.) vocabulary. It's not an English vocabulary. There's no default label to use. The obvious (and well-researched) solution is an entirely opaque, non-lexical URI. When I, wearing a developer hat, insist (as you do) that it makes the vocabularies virtually impossible to be used in development, my client regrets that there doesn't seem to be any other solution. The solution that we came up with was that, rather than have no lexical URIs, we would have _all_ of the lexical URIs, and declare them as owl:sameAs. We could have used owl:equivalentProperty (and we may have to in some cases where the translation isn't lexical but rather conceptual) but it's not as strong. The significant downside is that it immediately makes the vocabularies owl:full. At some point in the future, we may publish the mappings from lexical to opaque as a separate map for each language that can be included in the vocabulary or not and that would 'solve' the owl:full problem, sortof. The rejection of a single lexical URI in English wasn't 'politically correct' in the pejorative sense that we usually use that phrase, but rather an acknowledgement and embrace of a multilingual community. It was the politic thing to do. And yeah, our solution is debatable, but it's a debate we've often had over the years, with both colleagues and clients, in public and in private, and sometimes there's just no pleasing everyone, so we just do the best we can with the tools we have, eh? And build some new tools, which we're also working on. Cheers for the useful debate. Jon