Re: Re : Option 8 (was: Re: Service discovery redux - endpoint-based mechanisms)

On Aug 19, 2009, at 6:37 AM, Axel Polleres wrote:

> > If the service description if fetched from the exact endpoint URI,  
> and
> > the description looks like:
> >
> > <> a sparql:Endpoint ;
> >    sparql:has sparql:featureA ;
> >    ...
>
> If I understood correctly, the URI identifying the endpoint does not  
> necessarily need to be the URL of the endpoint, ie. cf. [1]
>
> Greg suggests in [1] the vocabulary providing a separate predicate
>
>  sd:url
>
> to point to the actual URL (in the case of your example probably the  
> public one https://machine.example.com/myapp/sparql/), so is that  
> really an issue?
>
> In other words, would the relative URL be really needed?

I don't think this is something we've talked about much, but I did  
that for two reasons:

1) Using sd:url is how DARQ did it, so there was an existing term to  
use.

2) I used a blank node for the service because coining an arbitrary  
URI seemed strange, but the access URL (where you get the service  
description, or an html form, or something else) clearly isn't the  
same resource as the service.

.greg

Received on Wednesday, 19 August 2009 18:00:41 UTC