Re: has, is, of

Hmm...

On Mon, 21 Apr 2014 15:10:59 +0200, Karen Coyle <kcoyle@kcoyle.net> wrote:

> Great links, thanks, and also for the insight into Japanese. It looks  
> like Pat Hayes has summed it up nicely in [2]:
>
> "one can always find nice intuitive evidence for whichever convention  
> you decide to like best."

Yes. And I note that Japanese and all the IndoEuropean languages I know  
including english would naturally provide grammatic indicators - although  
these would be associated to the subject or object in many cases.

X author Y is, in english grammar, ambiguous. Does it mean X is the author  
OfY, or XHas author Y? Or is it actually a symmetric property such as  
relatedTo?

Martin claims it is easy in practice. I think this is the case if there is  
a very clear convention such all properties assume that the Object (in RDF  
terms) should be considered a genitive/possessive (in grammatical terms),  
i.e. X something Y always means X is the something of Y, or the inverse (X  
something Y => X has a something which is Y).

But this assumption is undocumented, and until we sort out the issue of  
inverting property relations (do we add inverse properties for things, in  
which case they break whatever naming direction we chose and therefore  
need some special marker, or do we rely on the underlying syntax to allow  
this) we have some inverted properties.

Schema is not *that* small. And it isn't apparently in danger of shrinking.

It is used, with english documentation, by a large number of people whose  
grasp of the subtleties of english grammar is quite limited. Expecting  
them not to make mistakes means we should provide some very clear rules  
about interpretation.

As was also mentioned in the thread, I think we need to provide far  
clearer documentation of the meaning of terms - at the moment there is  
rarely enough to disambiguate anything - which is a problem in a language  
with as much ambiguity as english.

> It seems that my earlier suggestion -- that it may just come down to  
> being consistent within a given vocabulary -- is as close as we'll get  
> to agreement.

Any given vocabulary should aim to be consistent. I'd like to see them  
being really careful about that, especially if they start to grow.

cheers

Chaals

> kc
>
> On 4/21/14, 4:03 AM, KANZAKI Masahide wrote:
>> Hello,
>>
>> There has been discussions on whether property names should be nouns
>> or verbs. You might find JeniT's article [1] and RoleNoun entry at W3C
>> Wiki [2] interesting, for example.
>>
>> BTW, noun property names work better at least in Japanese, e.g.
>>
>>    :book :著者 :dan .      # 著者 = author
>>
>> seems OK, while verb form would be tricky: has-author style  
>> (:著者を持つ)
>> sounds strange partly because Japanese sentense has SOV structure
>> rather than SVO. Instead,
>>
>>    :book :の著者は :dan .  # ≒ whose author is
>>
>> makes sense, though almost not acceptable as a property name.
>>
>> cheers,
>>
>> [1] http://www.jenitennison.com/blog/node/128
>> [2] http://www.w3.org/wiki/RoleNoun
>>
>> 2014-04-21 7:12 GMT+09:00 Karen Coyle <kcoyle@kcoyle.net>:
>>>
>>>
>>> On 4/20/14, 12:11 PM, Dan Scott wrote:
>>>
>>>>
>>>> branchOf
>>>> causeOf
>>>> comprisedOf
>>>> estimatesRiskOf
>>>> increasesRiskOf
>>>> isPartOf
>>>> isVariantOf
>>>> memberOf
>>>> predecessorOf
>>>> successorOf
>>>>
>>>> And the currently used "isFoo" properties are:
>>>>
>>>> isAvailableGenerically
>>>> isBasedOnUrl
>>>> isConsumableFor
>>>> isFamilyFriendly
>>>> isGift
>>>> isPartOf
>>>> isProprietary
>>>> isRelatedTo
>>>> isSimilarTo
>>>> isVariantOf
>>>>
>>>
>>> I always get worried about language misunderstandings whenever  
>>> prepositions
>>> are involved. I don't know how all this reads to non-native speakers of
>>> either British or American English, but I do know that even between  
>>> those
>>> two the prepositions can vary: "Have a chat to" vs. "Have a chat with"  
>>> is
>>> pretty innocent vis-a-vis schema.org, but the American "agree to  
>>> something"
>>> is simply "agree something" in British English, so a property "agreeTo"
>>> would be strange to a British speaker. And I don't see what would be
>>> ambiguous about:
>>>    X -> related -> Y
>>> especially when read following the W3C document's model:
>>>
>>> Y is the value of -> related -> for X
>>> X has property -> related -> with a value Y
>>>
>>> although:
>>> the related -> of X is -> Y
>>>
>>> is awkward, whereas
>>> the title -> of X is Y
>>>
>>> is not. I agree with Thad's "KISS" - keeping it simple.
>>>
>>> kc
>>>
>>>
>>> --
>>> Karen Coyle
>>> kcoyle@kcoyle.net http://kcoyle.net
>>> m: 1-510-435-8234
>>> skype: kcoylenet
>>>
>>
>>
>>
>


-- 
Charles McCathie Nevile - Consultant (web standards) CTO Office, Yandex
       chaals@yandex-team.ru         Find more at http://yandex.com

Received on Monday, 21 April 2014 16:46:32 UTC