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

Re: Indicating main entity / primaryTopic - proposal to use 'schema.org/about'

From: Simon Spero <sesuncedu@gmail.com>
Date: Tue, 20 May 2014 14:45:04 -0400
Message-ID: <CADE8KM4WLT7JHQ+r2UuiVPrbhk5df1OoAOLJJeqw1k9gLJR4Dw@mail.gmail.com>
To: Dan Brickley <danbri@google.com>
Cc: W3C Web Schemas Task Force <public-vocabs@w3.org>
I ought to leave this to Karen, since Patrick Wilson was a professor at
Cal, but as Dan noted earlier,  "*the* main subject" of a work is
indeterminate.
See [1] for a snippet from "Two kinds of power".

It believe it is risky to interpret schema:sameAs (SSA) as having any
semantics beyond:

All x,y,z : SSA(x,z) & SSA(y,z) -> x = y

And that is dubious unless the full context [2] of both SSA assertions is
known (because SSA is defined in terms of the intentionality of the content
of a *deferenced* *URL*, which makes things less Cool).

[Speak not the name of TAG issue  2*7 thrice, lest it waken. ]

Simon

[1]
http://en.wikipedia.org/wiki/Subject_(documents)#Patrick_Wilson_.281927.E2.80.932003.29
[2] it's good to have a known context.  Or, talking about dereference, an
@context *cough*
On May 20, 2014 1:03 PM, "Dan Brickley" <danbri@google.com> wrote:

> On 20 May 2014 17:30, Jason Douglas <jasondouglas@google.com> wrote:
> > I think you buried the lede in the github link at the bottom.  :)
> ( https://gist.github.com/anonymous/cf7e24f6378b176aa010 )
>
> Ok, let's jump into the markup. It does flush out some issues -
>
>
> > In terms of markup would it be reasonable to do this?
>
> Slight tweaks (discussing hashes and slashes with Dan Scott and
> Stephane Corlosquet in IRC...):
>
>  <html vocab="http://schema.org/">
>    <head>
>      <link property="about" href="#main_item"/>
>    </head>
>    <body>
>      <div resource="#main_item" typeof="MusicEvent">
>        ...
>      </div>
>    </body>
>  </html>
>
> (this adds a '/' in the @vocab url, and # within @resource attribute)
>
> > and rely on an implicit WebPage subject?
>
> Looks fine from RDFa point of view. Even though "implicit WebPage" is
> a schema.org convention, generic RDFa 1.1 parsing would attach the
> 'about' property to the default base URL, which would be the page.
>
> As for Microdata I'm not sure the best pattern, but this with more
> RDFa-ish structure in the <head> might be bearable:
>
> <html>
>    <head>
>      <link rel="http://schema.org/about" href="#main_item"/>
>    </head>
>    <body>
>      <div id="main_item" itemscope itemtype="http://schema.org/MusicEvent
> ">
>        ...
>      </div>
>    </body>
>  </html>
>
>
> FWIW another everything in the body pattern that works for RDFa (but
> dips into RDFa 1.1 beyond the Lite subset) is:
>
> <div vocab="http://schema.org/" typeof="Person">
>  <link rev="about" href=""/>
>  <span property="name">John Smith</span>
> </div>
>
> Which brings me to another issue that's never far away - we have a lot
> of ways of saying quite similar but different things(*):
>
> <div vocab="http://schema.org/" typeof="Person">
>  <link rel="url sameAs officialWebPage" rev="about" href=""/>
>  <span property="name">John Smith</span>
> </div>
>
> At least this is pretty compact markup, but we owe the world an
> account of how at least 'url', 'sameAs' and 'about' interact.
>
> e.g. for any x, y where  x--sameAs-->y, is y--about-->x going to be true?
>
> sameAs's definition: "URL of a reference Web page that unambiguously
> indicates the item's identity. E.g. the URL of the item's Wikipedia
> page, Freebase page, or official website." ... which covers the
> official website use case.
>
>
> Dan
>
>
>
> (*) officialWebPage is used by Yandex e.g. see
> http://help.yandex.com/webmaster/supported-schemas/review-car.xml but
> similar ideas surfaced here in the social/profile/account discussion
>
>
Received on Tuesday, 20 May 2014 18:45:31 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:29:41 UTC