RE: Inverse Properties in Microdata:, was Re: schema.org update, v1.8: added WebSite type; broadened isPartOf to relate CreativeWorks

On 5 Aug 2014 at 15:06, martin.hepp@ebusiness-unibw.org wrote:
> Markus:
>> 
>>> Hence the desire for a convention on
>>> top of Microdata to match this (relatively niche and rarely used)
>>> piece of syntax that RDFa and JSON-LD publishers already have.
>> 
>> This is exactly the thing I me that triggered my question. If it is
really "relatively niche and
> rarely used", do we really want to complicate everyone's live by
introducing a new feature
> to Microdata? Or would it be more sensible to require those "few" who need
it to invest a
> little more effort?
> 
> The issue is: Without itemprop-reverse, there will be a strong demand to
> add more and more inverse variants of existing properties, which will
> increase the size of the vocabulary uneccessarily.

Or better examples which show how to mark this stuff up properly. Let's be
honest, most devs simply copy and paste examples to get their job done. This
is not a bad thing at all but it requires good examples which aren't
available for this stuff as Jarno pointed out.


> The original motivation of itemprop-reverse was that by adding such a
feature
> to Microdata we solve the problem in one place instead of at the level of
200+
> properties.

Is that really the only option to achieve that?


> Also, as Dan has already said nicely: Moving to RDFa can be a real
> burden of there are tons of existing site templates in Microdata. There
> are switching costs from one standard to another.

Yep, fully agreed.


> As a personal note, I think it is unfair to try to use ths proposal as a
> means for improving the fate of RDFa. The only valid criticism against
> the proposal I would accept are those from the formal mechanics of
> Microdata processing rules or from other aspects within Microdata. The
> market positions of competing syntaxes for structured data should not
> weigh in this decision.

I couldn't care less about the Microdata vs. RDFa "war". I know it sounds
like I would be a big proponent of RDFa. I'm not. Not at all. I just happen
to think that adding reverse properties to Microdata might not be the right
thing to do. Microdata's major selling point is simplicity. It achieves that
mostly by focusing on a single vocabulary and a tree-shaped data model.
Reverse properties break that model. It is also not clear how they could be
mapped to application/microdata+json as pointed out in another mail.


--
Markus Lanthaler
@markuslanthaler

Received on Tuesday, 5 August 2014 13:23:30 UTC