W3C home > Mailing lists > Public > public-vocabs@w3.org > January 2012

Re: Is name required in http://schema.org/person ?

From: Dan Brickley <danbri@danbri.org>
Date: Thu, 19 Jan 2012 11:22:59 +0100
Message-ID: <CAFNgM+agGkensLSQtUfbwSQbn_6eE7n8OpOk=eDxY_6cYiZtyw@mail.gmail.com>
To: Adrian Giurca <giurca@tu-cottbus.de>, Jocelyn Fournier <jocelyn.fournier@googlemail.com>
Cc: public-vocabs@w3.org
> On 1/19/2012 7:32 AM, Jocelyn Fournier wrote:
>> According to http://www.schema.org/Person, familyName & givenName could be
>> used in replacement for name.
>> However, if I'm not providing name, Rich Snippets Testing Tool complains
>> about missing required field "name (fn)".
>>
>> Is this a bug in the schema.org documentation, or in the Rich Snippets
>> Tool ?

It's a bit of both. I don't think """familyName: Family name. In the
U.S., the last name of an Person. This can be used along with
givenName instead of the Name property.""" is entirely adequate as a
definition (not least because the U.S. is very multi-cultural). But
ultimately, the schema.org documentation tells you more about the
meaning of the property, and less about its usage in some particular
(evolving, complex...) Web service.

So here, we should read this documentation as saying "schema.org's
givenName and familyName properties provide an alternative way of
describing a person's name to a general name property'. Whether
different parts of the Rich Snippets product support that is another
matter. Rich Snippets is not obliged as a schema.org implementation to
track all the equivalent or near-equivalent forms of expression
documented in schema.org. Also, schema.org is not obliged to (or
capable of) documenting all the shared information-consuming habits of
the services that use it. But that said, we can't be too abstract here
--- It seems reasonable to expect services to handle the most obvious
equivalencies, and for schema.org to document the most obvious very
general-purpose common patterns.

On 19 January 2012 09:52, Adrian Giurca <giurca@tu-cottbus.de> wrote:
> I would say, actually Schema.org does not define required properties. But
> maybe I'm wrong...

At its core, schema.org is more or less a kind of dictionary, in that
it describes the meaning of some terms you can use. Just as
dictionaries don't tell you exactly what to say, nor does schema.org.
Specific products and services (like Rich Snippets) will have specific
information needs and expectations, and ultimately it's their job to
communicate those details. However in practice there are some patterns
that are fairly common across applications. The high-level properties
that apply to everything (ie. to http://schema.org/Thing), i.e.
description, image, name, URL (alongside a type), are generically
useful in many schema.org Web-based apps. You see similar in many
contexts, for example Open Graph Protocol has title/type/url/image at
the top level. Beyond this, each class of thing described at
schema.org is described with a bundle of properties, and so (stating
the obvious?) many applications often expect a few of those properties
together to do something useful.  It is unusual for example for an
application that does something useful with 'longitude' to not
simultaneously require 'latitude' (and also sometimes 'elevation'...).
But we cannot attempt to catalogue all such applications and their
needs, which is why schema.org's data model and schema system can seem
strangely passive and declarative compared for e.g. to XML DTDs, which
are much more 'needy'.

Schema.org is fundamentally a collaboration around vocabulary and
shared data structures, rather than around products or applications,
so its main documentation is much more focussed on the terms and some
example descriptions than on 'musts' and 'shoulds' regarding what
information is needed in specific application-specific contexts. This
emphasis allows us to flow information between a wider variety of
environments. A dictionary-like approach, on its own, can feel
frustratingly under-specified. We have a space in W3C's wiki for more
informal, additional information - http://www.w3.org/wiki/WebSchemas -
that can be used to track more specific constraints, application
information needs etc. (or links to those, e.g. to Rich Snippets
documentation or docs from other projects using schema.org
vocabulary).

As for whether to report Rich Snippets issues here --- I'll try to
find a more appropriate and direct mechanism (and I'll do my best to
relay issues to that team), but for now it is good to hear back from
people working with practical markup details of schema.org deployment,
so mailing here is fine. It's best not to assume that mail sent here
will necessarily be read by the Rich Snippets team but the topic is
perfectly in scope for this list to discuss.

Hope that helps,

Dan
Received on Thursday, 19 January 2012 10:23:29 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 22 May 2012 06:48:59 GMT