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

Re: Socialnetworks of a person or organization

From: <martin.hepp@ebusiness-unibw.org>
Date: Thu, 10 Apr 2014 11:53:39 +0200
Cc: Jason Douglas <jasondouglas@google.com>, Dan Brickley <danbri@google.com>, David Deering <david@touchpointdigital.net>, W3C Web Schemas Task Force <public-vocabs@w3.org>
Message-Id: <361F261D-9C44-4710-8EE7-F06DE6A2499D@ebusiness-unibw.org>
To: Charles McCathie Nevile <chaals@yandex-team.ru>
On 10 Apr 2014, at 09:08, Charles McCathie Nevile <chaals@yandex-team.ru> wrote:

> On Wed, 09 Apr 2014 17:46:00 +0200, Dan Brickley <danbri@google.com> wrote:
> 
>> On 9 April 2014 16:24, Jason Douglas <jasondouglas@google.com> wrote:
>>> -1  There's a difference between reference pages *about* the same entity and pages authored/controlled *by* the same entity.
>> 
>> Is it a difference we want to fully capture here?
> 
> IMHO yes, but


I personally suspect that the distinction between 

1. entities and 
2. Web resources about the entities 

does not have to be exposed at markup level but can be inferred from context and other signals by a search engine or another client.

When you say that a certain song is written by http://en.wikipedia.org/wiki/John_Lennon and or that you offer a trip to http://en.wikipedia.org/wiki/Yushan_(mountain), it is easy to infer for a computational service that not the Wikipedia Web page wrote the song, and that you do not want to climb upon a pile of HTML.

HttpRange-14 and related discussions are overrated, in my opinion. For broad audiences of Webmasters, it is much easier to provide the string of an authoritative Web page URI as a pointer to support entity recognition than to understand the difference between e.g.

http://dbpedia.org/page/John_Lennon 

and

http://dbpedia.org/resource/John_Lennon

In particular, if content negotiation will always make the "page" URI appear in a browser's address line, so for copy-and-paste, you need extra  effort to use the proper entity URI.

By the way, schema.org as a whole decided back then, and I think it was a great decision, to use the same URI for the page describing a conceptual element and the conceptual element itself. http://schema.org/Vehicle for instance is the identifier for the class/type and also a Web page.

Again, I urge you all to regard schema.org as an interface to Web developers' minds and the data they control, not as a beautiful all-purpose category system for the human race as a whole.

Martin


On 10 Apr 2014, at 09:08, Charles McCathie Nevile <chaals@yandex-team.ru> wrote:

> On Wed, 09 Apr 2014 17:46:00 +0200, Dan Brickley <danbri@google.com> wrote:
> 
>> On 9 April 2014 16:24, Jason Douglas <jasondouglas@google.com> wrote:
>>> -1  There's a difference between reference pages *about* the same entity and pages authored/controlled *by* the same entity.
>> 
>> Is it a difference we want to fully capture here?
> 
> IMHO yes, but
> 
>> I also control my homepage and various other pages that are not my
>> socialAccount.
> 
> In what sense is your "homepage" not a social account, and is *that*
> a difference we want to capture?
> 
>> But I've just heard another problem which I think also counts against
>> this proposal: some entities (e.g. news organizations) have dozens of
>> e.g. Twitter accounts ('sports news', 'music' etc). While it might be
>> reasonable to point to them all with e.g. socialAccount, it might be
>> that the best reference page for the entity is something like
>> http://en.wikipedia.org/wiki/BBC --- and they'll all end up smushed
>> together in a confusing way.
> 
> Hmm. If you want to *identify* yourself with a page (your homepage), a
> wikipedia entry, your twitter account, then you are schema:sameAs, no?
> 
> But if you want to say that you are the author of a page, such as an
> article you wrote or something that you contribute to, you'd do
> something different, analagous to performerIn, that isn't yet nicely
> documented but is the inverse of the "author" property.
> 
>> So given Jason's point and this observation I'll back off from the
>> proposal. So much for thinking-out-loud.
> 
> (I am still doing it here…)
> 
>> Maybe the core concept is 'account', which suggests an account holder
>> and a service provider, and hints at the ability to show (openid
>> connect etc.) that you're the account holder.
> 
> I don't think that is particularly core here, unless you're using the account as an ID. And then you probably really want something that says that explicitly (and to think 3 more times before we go there).
> 
> The identifying use case for social networks can be met by people with their own domain (no relevant service provider, no openID, …). It's just an "about" page. Foaf used seeAlso for this, as a hack that pointed to things which were worth crawling - maybe that wasn't such a bad idea.
> 
>> Dropping the word 'social' (which was discussed here a while back)
>> does remove some of the fuzzyness.
> 
> Yes.
> 
>> <div itemscope itemtype="http://schema.org/Person">
>> <span itemprop="name">Stephen Fry</span>
>>    (<a itemprop="url" href="http://www.stephenfry.com/">stephenfry.com</a>,
>>     <a itemprop="account" href="http://twitter.com/stephenfry">twitter</a>,
>>     <a itemprop="sameAs"
>> href="http://en.wikipedia.org/wiki/Stephen_Fry">wikipedia</a>)
>> </div>
> 
> Hmm. Food for thought.
> 
> cheers
> 
> 
> -- 
> Charles McCathie Nevile - Consultant (web standards) CTO Office, Yandex
>      chaals@yandex-team.ru         Find more at http://yandex.com
> 
Received on Thursday, 10 April 2014 09:54:09 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:29:39 UTC