Re: Call for consensus on defining IRI template expansion (ISSUE-30)

> If it's a boolean with a default value of false you just need to check for
> true. You also know that there will only ever be two ways to do this, not
> dozens of them.

There could also be a default for a non-boolean property.
(And note that "defaults" are hard in an open world anyway;
 but we can indeed make assumptions.)

> I don't consider being able to describe arbitrary
> existing Web APIs with Hydra an important requirement anymore. As soon as
> you start describing requests that involve payloads, you will run into lots
> of issues that aren't worth the hassle and the required complexity to make
> them work IMO.

Sure, but keeping our options open in the future seems reasonable.

> Interoperability is created my eliminating
> variability as much as possible. I think this is a low-hanging fruit.

Mmm, also true…

But then I'd again say that a non-boolean is not more complicated;
so we get possible future extensibility at no cost.

>>>>>> - the datatype xsd:string is always omitted
>> 
>> It's just a datatype like any other. I would not disallow it for that
>> reason.
> 
> You are probably right. But then even the SHOULD doesn't make much sense. In
> that case, a MAY omit is probably better.

+1

Ruben

Received on Thursday, 24 July 2014 07:26:29 UTC