W3C home > Mailing lists > Public > public-vocabs@w3.org > September 2014

Re: ItemList proposal

From: ☮ elf Pavlik ☮ <perpetual-tripper@wwelves.org>
Date: Wed, 17 Sep 2014 15:50:56 +0200
Message-ID: <541991C0.2080503@wwelves.org>
To: "martin.hepp@ebusiness-unibw.org >> Martin Hepp" <martin.hepp@ebusiness-unibw.org>
CC: public-vocabs@w3.org
On 09/17/2014 01:55 PM, martin.hepp@ebusiness-unibw.org wrote:
> On 16 Sep 2014, at 23:10, ☮ elf Pavlik ☮ <perpetual-tripper@wwelves.org> wrote:
>> On 09/12/2014 02:56 PM, martin.hepp@ebusiness-unibw.org wrote:
>>> So people can use schema:Ascending or schema:Descending while still being able to use free text for other cases?
>>> For "unordered", I would simply omit the itemListOrder property.
>> Doesn't it put us in Closed-world assumption?
> I do not think so, because in this particular case, CWA and OWA lead to the same effect - unordered and an unknown order seem to be equivalent to me. All unordered things may be ordered by an unknown ordering relation.
>> I must admit not knowing if schema.org tries to follow Open-world
>> assumption.
> As for schema.org: First of all, OWA vs. CWA is typically not decided on at the level of the the vocabulary, though it may influence modeling choices in the vocabulary.
> I think it is safe to assume that the sponsors of schema.org do neither adhere to strict OWA, but also not to strict Closed-Word Assumption based on Negation-as-Failure [1]. The absence of a statement does not imply its falseness, but it often tells us something.
> If a schema:Offer indicates no schema:DeliveryChargeSpecification, it is a bit more likely that there are none than that they are just not exposed. If you have opening hours indicated for Mondays and Wednesdays but nothing else, it is likely that the business is closed on the other days of the week.
> Properly interpreting the absence of a statement is very challenging and more complex that straight CWA or OWA. Search engines will likely use a lot of contextual information for this task. As a first simple heuristic, you can try to extract the part of the vocabulary actively used in a data source (e.g. Web site). If there a certain type or property is used on a number of entitities, then the absence of type information or a property is more significant than the absence of other types or properties.
> A somewhat related paper is
>     Polleres/Feier/Harth: Rules with contextually scoped negation, ESWC 2006, 
>     http://dl.acm.org/citation.cfm?id=2094646
> Best wishes / Mit freundlichen Grüßen
> Martin
Thanks Martin for this very helpful comment!
Received on Wednesday, 17 September 2014 13:53:17 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:29:44 UTC