Re: Shapes vs Classes (in LDOM)

On 2/6/15 6:31 AM, Arthur Ryman wrote:
> Holger Knublauch <holger@topquadrant.com> wrote on 01/29/2015 06:55:33 PM:
>
>> BTW wouldn't it have been better to use "oslc:shape" instead of
>> "oslc:instanceShape", because you don't really talk about "instances"
>> (of classes)?
> Holger,
>
> Historically, we first defined oslc:resourceShape as a way to link a shape
> with a service endpoint URI (essential a container) where you could create
> resources via POST or query them via GET.
>
> However, we needed a way to support PUT operations. When a client GETs a
> resource it can look in it for an oslc:instanceShape triple. That tells
> the client how it can modify the resource and PUT it back to the server.
>
> The point is that if we just used one oslc:shape property then its meaning
> would be ambiguous in the RDF description of a service endpoint. A service
> endpoint description document needs to refer to its own shape as an RDF
> graph AND the shape of the RDF graphs that you, e.g., POST to it.

My suggestion was to simply rename :instanceShape to :shape, not to 
merge this property with another property. I may not have been clear, or 
did oslc:shape already have another meaning?

>
> The term "instance" was used because in many cases the RDF graph does in
> fact represent some instance of an OO programming language object that
> lives in the server. We picked the term so that typical web application
> developers would understand it.

Yep, that's the same reason why I believe sticking to general OO would 
have benefits.

Holger

Received on Thursday, 5 February 2015 22:31:24 UTC