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

Re: Schema.org v3.3 release candidate for review

From: Vicki Tardif Holland <vtardif@google.com>
Date: Wed, 24 May 2017 14:40:54 -0400
Message-ID: <CAOr1obFFkFiQnm=UnLj1TmK3sFq06RzN6J7nGR16xCa5xuiUsA@mail.gmail.com>
To: Richard Wallis <rjw@dataliberate.com>
Cc: Thad Guidry <thadguidry@gmail.com>, Martin Hepp <mfhepp@gmail.com>, Richard Wallis <richard.wallis@dataliberate.com>, "R.V.Guha" <guha@guha.com>, Nicolas Torzec <torzecn@yahoo-inc.com>, "schema.org Mailing List" <public-schemaorg@w3.org>, St├ęphane Corlosquet <scorlosquet@gmail.com>, Dan Brickley <danbri@google.com>, Charles McCathie Nevile <chaals@yandex-team.ru>, Tom Marsh <tmarsh@exchange.microsoft.com>
>
> And how does one say it is not a public place?


If we really wanted to know that explicitly, we could create a new type.

Even with a boolean, most folks will omit the "default" value, and the
default is going to switch between true and false depending on context, so
data readers are still left with the open world problem even if we think we
have engineered a work around.

- Vicki


On Wed, May 24, 2017 at 1:29 PM, Richard Wallis <rjw@dataliberate.com>
wrote:

> And how does one say it is not a public place?
>
> ~Richard
>
>
> On 24 May 2017, at 18:35, Vicki Tardif Holland <vtardif@google.com> wrote:
>
> In this case, having a property to flag if a Place is accessible by public
>> visitors covers more ground than a Type AND is easier for publishers.
>
>
> I don't follow. If they use multiple types, they can say it is a public
> place and a park.
>
> And a boolean does not allow places like King's Chapel in Boston, which is
> often publicly accessible, but not during church services.
>
> - Vicki
>
>
> On Wed, May 24, 2017 at 10:19 AM, Thad Guidry <thadguidry@gmail.com>
> wrote:
>
>> I did some digging and scenarios of True and False on this new
>> publicAccess property on Place across some atypical Places.
>> In the case of a boolean for "publicAccess" ...
>>
>> We have Park under CivicStructure but that's not always the case...Not
>> all Parks are actually publicly accessible or even public, some are
>> actually private but still name themselves a park.  Example of a famous one
>> in New York City: https://www.google.com/search?q=gramercy+park+new+york
>>
>> In this case, having a property to flag if a Place is accessible by
>> public visitors covers more ground than a Type AND is easier for publishers.
>>
>> Backtracking and agreeing with Martin and Richard on this particular
>> property of publicAccess.
>>
>> -Thad
>>
>> On Wed, May 24, 2017 at 6:34 AM Martin Hepp <mfhepp@gmail.com> wrote:
>>
>>> I might miss the point, but I have a few concerns:
>>>
>>> 1. Substituting Boolean properties by types will work only if we have
>>> full support for multi-typed entities in the major search engines. As soon
>>> as there are adverse effects of making an entity multi-typed, we cannot
>>> substitute a Boolean property by a new type.
>>>
>>> 2. Also, Boolean properties, like faceted classifications, allow us to
>>> classify an object along multiple dimensions. As soon as we have a subclass
>>> hierarchy, using types can quickly create at least confusion but often
>>> inconsistencies.
>>>
>>> 3. From a theoretical perspective, qualitative properties and even
>>> quantitative properties can also create a secondary type system.
>>>
>>> So in a nutshell, I think Boolean properties have their right if we want
>>> to add a distinction or categorial information without messing with the
>>> type hierarchy of the main type.
>>>
>>> Martin
>>> -----------------------------------
>>> martin hepp  http://www.heppnetz.de
>>> mhepp@computer.org          @mfhepp
>>>
>>>
>>>
>>>
>>> > On 24 May 2017, at 13:24, Richard Wallis <
>>> richard.wallis@dataliberate.com> wrote:
>>> >
>>> > In most cases I agree with you.
>>> >
>>> > However in this case the boolean property was proposed to enable not
>>> only the definition that a Place is open for publicAccess, but also a Place
>>> is not open for publicAccess.
>>> >
>>> > This came from the enhancements to TouristAttraction proposals where
>>> many places may well be still of interest regardless of if public access is
>>> available or not; whilst that accessibility is still useful information.
>>> Following the logic of defining a PublicPlace, would lead in this case to
>>> creating a NonPublicPlace type to enable that capability which I believe is
>>> even more clunky than the proposed boolean.
>>> >
>>> > ~Richard.
>>> >
>>> >
>>> >
>>> > Richard Wallis
>>> > Founder, Data Liberate
>>> > http://dataliberate.com
>>> > Linkedin: http://www.linkedin.com/in/richardwallis
>>> > Twitter: @rjw
>>> >
>>> > On 22 May 2017 at 19:05, R.V.Guha <guha@guha.com> wrote:
>>> > I agree. I prefer types
>>> >
>>> > On May 22, 2017 10:55 AM, "Vicki Tardif Holland" <vtardif@google.com>
>>> wrote:
>>> > We should figure out a principled approach to boolean properties. I am
>>> not a fan of them as they create a secondary type system (publicAccess
>>> could also be PublicPlace), but because they are not actually types, you
>>> cannot add properties to them. For example, you cannot say when the public
>>> access hours are if they differ from other hours.
>>> >
>>> > With that said, it is probably not worth holding up the release.
>>> Otherwise, LGTM.
>>> >
>>> > - Vicki
>>> >
>>> > On Mon, May 22, 2017 at 1:20 PM, Dan Brickley <danbri@google.com>
>>> wrote:
>>> >
>>> >
>>> > On 22 May 2017 at 18:11, Chaals is Charles McCathie Nevile <
>>> chaals@yandex-team.ru> wrote:
>>> > I already  made some comments on HowTo.
>>> >
>>> > Thanks - sensible tweaks, we should fold those in.
>>> >
>>> > I'm not enamoured of filling up on reverse properties - as far as I
>>> can tell they are only for microdata, and I'm not sure why people couldn't
>>> just use RDFa Lite instead, if microdata isn't serving their purposes -
>>> which I suspect for many interesting cases it doesn't.
>>> >
>>> > There is some ongoing discussion of that here -
>>> https://github.com/schemaorg/schemaorg/issues/1156 - and an agreement
>>> to revisit the reverse properties before any move from Pending into a named
>>> extension area (or the core).
>>> >
>>> > Otherwise, LGTM, please go ahead.
>>> >
>>> > Thanks!
>>> >
>>> > cheers
>>> >
>>> >
>>> >
>>> > On 22/05/17 18:06, Dan Brickley wrote:
>>> > Dear Schema.org Community Group, Steering Group,
>>> >
>>> > Based on our consensus discussions here and in Github, here is a
>>> > proposal for a new Schema.org release, version 3.3:
>>> >
>>> > http://webschemas.org/docs/releases.html#v3.3
>>> >
>>> > I'd like to aim at publishing this around June 5th. Bugs, mistakes,
>>> > typos, modeling and example improvements and other detailed review
>>> > comments are welcome here or in the issue tracker at
>>> > https://github.com/schemaorg/schemaorg/issues/1569
>>> >
>>> > cheers,
>>> >
>>> > Dan
>>> >
>>> > ps. as usual there are a few pieces of the release that will be put
>>> together
>>> > at the end (anything involving exact release dates, dated snapshots
>>> etc.).
>>> >
>>> > --
>>> > Charles McCathie Nevile   -   standards   -   Yandex
>>> > chaals@yandex-team.ru - Find more at http://yandex.com
>>> >
>>> >
>>> >
>>> >
>>> >
>>>
>>>
>>>
>
Received on Wednesday, 24 May 2017 18:41:30 UTC

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