Re: property/class ambiguity in languages with no letter case

Hi Bernard,
Thanks - I did not see that when writing my reply (but the issue also comes up in the context of schema.org regularly). As for “religious”, I meant that as “with a strong mental bond to the subject”, “eagerly looking for a single truth”.
;-)
Martin


On 12 Feb 2014, at 08:17, Bernard Vatant <bernard.vatant@mondeca.com> wrote:

> Hi Martin
> 
> Just for the record, the initial question by Phil was not about naming in the schema.org namespace, but vocabularies hosted in the W3C ns, such as org. This is public-vocabS list ...
> And my point about Chinese was indeed this language seems to have less religion and more built-in pragmatism than English - but maybe your "religious" remark was not addressed to me after all :)
> 
> 
> 
> 2014-02-11 18:00 GMT+01:00 martin.hepp@ebusiness-unibw.org <martin.hepp@ebusiness-unibw.org>:
> A general comment:
> 
> When we articulate requirements on the naming of elements in schema.org, let’s 
> 
> 1. not get too philosophical and 
> 2. look at how keywords have been chosen in other formalisms, namely programming languages.
> 
> Of course, it is more difficult to find catchy keywords for a broad conceptual schema that for the set of instructions in a programming language. But on the other hand I find the implicit requirement of an “ideal” grounding of schema.org in various human languages too far reaching.
> 
> Python, Java, etc. and most programming languages except for machine code have dealt pretty well with mostly English-based keywords, and have been used successfully by large, diverse audiences in multi-national development teams.
> 
> For instance, “print” in many languages from BASIC to Python is incorrect, when compared to the etymology of the word, see http://www.etymonline.com/index.php?term=print.
> 
> So IMHO, let’s not get too religious about naming.
> 
> Best
> 
> Martin
> 
> PS: Some people have thought about the issue from the perspective of mapping Chinese database element names, see http://dl.acm.org/citation.cfm?id=1141673.
> 
> 
> 
> 
>  
> On 11 Feb 2014, at 15:03, Jindřich Mynarz <mynarzjindrich@gmail.com> wrote:
> 
>> Hi,
>> 
>> I think that in most cases foo -> Foo (i.e. property foo having rdfs:range instances of class Foo) can be read as 'has Foo'. However, as Dan pointed out, there are some cases, for which this doesn't hold. It's not a convention everyone follows.
>> 
>> I wanted to have a broader look at this pattern, in which properties use lower-cased local name of classes in their range. So I queried LOV SPARQL endpoint (http://lov.okfn.org/endpoint/lov), downloaded vocabularies it lists (those that resolved), and queried the merged dataset to see what properties and classes following the mentioned naming pattern are there.
>> 
>> This sample contains 251 such property-class pairs and it's available at http://databin.pudo.org/t/289185. The (hidden) third column contains "yes" if the triple property rdfs:range class is present.
>> 
>> I hope this sample can serve as an approximate illustration of this widespread naming pattern in practice.
>> 
>> Best,
>> 
>> Jindřich
>> 
>> -- 
>> http://mynarz.net/#jindrich
>> 
>> 
>> On Tue, Feb 11, 2014 at 2:18 PM, Dan Brickley <danbri@google.com> wrote:
>> 
>> On 11 Feb 2014 03:03, "Phil Archer" <phila@w3.org> wrote:
>> >
>> > On this occasion I really am trying to avoid getting into the debate about whether it is right or not to use an object property with a label that is the same as the class that is its range, differentiated only by the case of the first letter. That is an issue, and we prob should clear it up, but not today (and I suspect there is a lot of agreement on this).
>> >
>> > I'm just asking, do you agree or not that foo -> Foo *implies* 'has foo' -> Foo sufficiently strongly that a translation of the label into a language that does not have upper and lower case letters can indeed be 'has foo?'
>> Good point re labels. 
>> 
>> Some examples where this doesn't work well:
>> 
>> subClassOf -> hasSubClassOf
>> alumniOf -> hasAlumniOf
>> 
>> The 1st points to a broader type; the 2nd to an Educational org that someone is an alumni of.
>> 
>> In both cases, prepending 'has' makes less sense and encourages the property to be misread backwards (until you read the last syllable and get confused).
>> 
>> In general 'xyzOf' seems to me a 'last resort' when looking for property names. I wish we'd called rdfs:subClassOf "superClass" instead. But as you say that's another discussion.
>> 
>> Some more potential examples - https://dvcs.w3.org/hg/webschema/file/861b6fd28fcb/schema.org/translations/zhcn/schema_org_zhcn.html
>> 
>> This is a first cut at schema.org labels in Chinese, thanks to Baidu, where lang of zh-cn is "Mainland China, simplified characters", and as I understand it, caseless.
>> 
>> First, a class, 'Event':
>> 
>>      <div typeof="rdfs:Class" resource="http://schema.org/Event
>> ">
>> 
>>    
>>        <span class="h" property="rdfs:label">Event</span>
>> 
>>    
>>        <span class="h" property="rdfs:label" xml:lang="zh-cn">事件</span>
>> 
>>    
>>        <span property="rdfs:comment">An event happening at a certain time at a certain location.</span>
>> 
>>    
>>        <span property="rdfs:comment" xml:lang="zh-cn">在某时某地发生的一件事</span>
>> 
>>            <span>Subclass of: <a property="rdfs:subClassOf" href="http://schema.org/Thing
>> ">Thing</a></span>
>> 
>>         </div>
>> and a property, 'event', which seems to have an identical zh-cn label:
>> 
>> 
>>       <div typeof="rdf:Property" resource="http://schema.org/event
>> ">
>> 
>>  
>>        <span class="h" property="rdfs:label">event</span>
>> 
>>  
>>        <span class="h" property="rdfs:label" xml:lang="zh-cn">事件</span>
>> 
>>  
>>        <span property="rdfs:comment">Upcoming or past event associated with this place or organization.</span>
>> 
>>  
>>        <span property="rdfs:comment" xml:lang="zh-cn">即将发生的或已经发生的跟该地点或组织有关的事件</span>
>> 
>>         <span>Domain: <a property="http://schema.org/domain" href="http://schema.org/Organization
>> ">Organization</a></span>
>> 
>>         <span>Domain: <a property="http://schema.org/domain" href="http://schema.org/Place
>> ">Place</a></span>
>> 
>>         <span>Range: <a property="http://schema.org/range" href="http://schema.org/Event
>> ">Event</a></span>
>> 
>>       </div>
>> Not sure these chars and markup will make it through email fully to everyone, but see webschemas mercurial repo at https://dvcs.w3.org/hg/webschema/file/861b6fd28fcb/schema.org/translations/zhcn/schema_org_zhcn.html for more examples. 
>> 
>> 
>> 
>> Dan
>> 
>> > Phil.
>> >
>> >
>> > On 11/02/2014 10:46, Jindřich Mynarz wrote:
>> >>
>> >> OK, I thought I must have misunderstood that. (However, you can argue that
>> >> you can provide owl:equivalentProperty links between the translated URIs.)
>> >>
>> >> If translating rdfs:labels is indeed the case, then why not have 2
>> >> vocabulary terms with the same label? Is it because it confuses vocabulary
>> >> users and worsens usability of the vocabulary in question? What other
>> >> concerns do you have on mind?
>> >>
>> >> - Jindřich
>> >>
>> >
>> > -- 
>> >
>> >
>> > Phil Archer
>> > W3C Data Activity Lead
>> > http://www.w3.org/2013/data/
>> >
>> > http://philarcher.org
>> > +44 (0)7887 767755
>> > @philarcher1
>> >
>> 
> 
> 
> 
> 
> -- 
> Bernard Vatant
> Vocabularies & Data Engineering
> Tel :  + 33 (0)9 71 48 84 59
> Skype : bernard.vatant
> http://google.com/+BernardVatant
> --------------------------------------------------------
> Mondeca                             
> 3 cité Nollez 75018 Paris, France
> www.mondeca.com
> Follow us on Twitter : @mondecanews
> ----------------------------------------------------------

Received on Wednesday, 12 February 2014 11:07:01 UTC