W3C home > Mailing lists > Public > public-schemaorg@w3.org > May 2017

Re: Does WaPo Fact Checker get the semantics of schema.org's ClaimReview wrong?

From: Dan Brickley <danbri@google.com>
Date: Tue, 9 May 2017 18:21:07 +0100
Message-ID: <CAK-qy=7maRrEa+Exsyy4LYYX8TtNHwvUTyk3pmUxMeNWaCYbng@mail.gmail.com>
To: Tim Finin <finin@umbc.edu>
Cc: Aaron Bradley <aaranged@gmail.com>, "schema.org Mailing List" <public-schemaorg@w3.org>
Thanks, both. The problem here was that when the fact checking
collaboration emerged last year, we realised that the schema.org
markup could be really minimalistic since the fact checkers were
mapping their heterogenous schemes to the numeric scale. It looks like
schema.org's http://schema.org/Rating documentation under-states the
extent to which "better = higher"; "A rating is an evaluation on a
numeric scale, such as 1 to 5 stars." doesn't capture the idea, so
unless you notice the bestRating ("The highest value allowed in this
rating system. If bestRating is omitted, 5 is assumed.") and
worstRating properties, you might reasonably assume they're
unconstrained. So there's a general issue with "Rating", some
under-documentation around the ClaimReview markup, and perhaps it is
worth coming up with some construction to represent simple numeric
rating scales more explicitly, and that allows the mapping to schemes
where higher = worse to be written down. It's 21 years since
https://www.w3.org/TR/REC-PICS-services-961031#Rating System after
all...

I'll get this reflected into a Github issue or two...

cheers

Dan


On 8 May 2017 at 20:34, Tim Finin <finin@umbc.edu> wrote:
> I just checked a few recent politifact items and they do indeed have the
> scale inverted.  :-(
>
> On Mon, May 8, 2017 at 12:48 PM, Aaron Bradley <aaranged@gmail.com> wrote:
>>
>> I haven't re-checked since I tweeted this on 11 April, but at that time
>> PolitiFact seemed also was using a scale that was the inverse of the
>> schema's specs:
>> https://twitter.com/aaranged/status/851902710307737600
Received on Tuesday, 9 May 2017 17:21:42 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:12:35 UTC