- From: Garret Wilson <garret@globalmentor.com>
- Date: Thu, 18 Oct 2007 17:53:28 -0200
- To: Bruce D'Arcus <bdarcus@gmail.com>
- CC: Sandro Hawke <sandro@w3.org>, semantic-web@w3.org
Bruce D'Arcus wrote: > >>> Question: are the difficulties around these issues in RDF a >>> consequence of the open world assumption, and the need to be able to >>> merge statements from disparate graphs? >> >> No, I don't think that's the core issue here. The issue is simply >> that there are many types of relationships in the universe, whether >> it's open or closed, that cannot easily be represented by RDF because >> RDF doesn't allow properties to be ordered. Needing to order >> properties without resorting to heavy list-list classes is a *very* >> common use-case. > > Right, but I'm not clear how you merge ordered lists? That's certainly a valid issue, but in the real world I don't think it will cause many problems. I believe that in the most common use cases, the data that needs to be ordered will be presented at the same time---ordering usually relates values, so ordered properties will usually be presented together. If it is the case that two data sets with ordered properties are merged (e.g. ordered ex:lastName from data set one, ordered ex:lastName from data set two), there are multiple ways to handle it. Even the naive way of ignoring order and merging properties doesn't wind up with too bad of a solution: the ordered properties are still in order *relative to each other*, although (via duplicated scoped order properties) they are intermingled with the other ordered properties. In other words, merging orders [0, 1, 2] and [0, 1, 2] produces orders [0, 0, 1, 1, 2, 2]. A more thoughtful merging would simply advance the order values of one of the sets so that the properties remain with the same relative order, but simply one ordered property set comes after the other. That is, merging orders [0, 1, 2] and [0, 1, 2] produces orders [0, 1, 2, 3, 4, 5]. Each of these scenarios I can easily live with. However, if you want to merge two values that are rdf:Lists, don't you wind up with the same issues? Sure, you can naively add properties and have one list of [A, B], another of [C], another of [D], another of [E, F, G]. That seems bulkier to me, though. > E.g. the unordered nature of RDF is not some dumb hack, but a design > feature. Just to be clear---I don't think the unordered nature of RDF is a dumb hack, either. But it has some shortcomings, as I've pointed out. > > We've argued about this before, but to summarize, I think your way of > describing this problem is a function of your programming background It's possible. But you make it sound like that's a bad thing. :) > and does not reflect real world name usage. Ah, that's a separate issue altogether. I'd beg to differ, but who can say who's right? No one has had the option in RDF, so of course no one is going to use it. ;) We'll have to wait and see how useful it proves to be. It doesn't cost anything to try URF. I have personally tried it, and I like it. :) > > I think I'll be content when SPARQL supports reasonable query of lists. Fine. Maybe it's not a problem you have. This is a problem I've repeatedly had, and I'm happy now. It seemed to be a problem Sandro was having, too. Best, Garret
Received on Thursday, 18 October 2007 19:54:51 UTC