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

Re: ItemList proposal

From: Martin Hepp <martin.hepp@unibw.de>
Date: Fri, 12 Sep 2014 16:08:03 +0200
Cc: Peter Mika <pmika@yahoo-inc.com>, "Jason Johnson (BING)" <jasjoh@microsoft.com>, W3C Web Schemas Task Force <public-vocabs@w3.org>
Message-Id: <2EB2D653-F044-4E87-966F-A265C3400F60@unibw.de>
To: Vicki Tardif Holland <vtardif@google.com>

On 12 Sep 2014, at 15:37, Vicki Tardif Holland <vtardif@google.com> wrote:

> 
>  
> -- itemListElement currently has the range Thing or ListItem or Text, which is triple redundant ;) First, we always accept Text instead of Thing, and ListItem is a type of Thing. So the range is essentially Thing?


Even if Thing logically includes schema:Text, I strongly recommend to use Thing or ListItem or Text, or to upddate the schema.org meta-model so that you can give hints on recommended subtypes if the formal type for a range is a very broad one.
> 
> 
> Yes. We were looking for a way to signal that there is a new class ListItem, which allows authors to express more than the item. Perhaps this is a suggestion we need better documentation.

I think the essential thing is that these choice of types is explicitly listed in the HTML documentation. In the current toolchain, this can be triggered with no effort and no negative side-effects by the logically redundant list if Thing or ListItem or Text. Also, since schema:rangeIncludes is pretty fuzzy, there is, IMO, no real logical conflict.

>  
> -- previousItem/nextItem has range ListItem. How do I capture the end of list? Is there a distinguished  NULL value?
> 
> I was imagining it working like a linked list. If a ListItem has no nextItem, it is the end of the list.

+1

Martin
Received on Friday, 12 September 2014 14:08:59 UTC

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