Re: Issue-71: the first bug tracking example

Alexandre Bertails <bertails@w3.org> wrote on 05/29/2013 06:18:11 AM:

> ...
> > If we don't have membershipPredicate in LDP 1.0 there is no way we can
> > add it later without breaking backwards compability. A server would 
have
> > to keep using rdf:member or break clients that expect it.
> 
> I don't understand why adding it later would break anything. It would be 

> a plain conservative extension, with backward compatibility.

If a client expects to be able to find the member resources of a container 
by looking for the predicate rdf:member there is no way we can change that 
later without breaking that client, is there?
We can't just later decide that servers may choose to use a different 
predicate and simply specify it with ldp:membershipPredicate. A client 
that is implemented against LDP 1.0 would look for rdf:member and wrongly 
conclude that the container has no members.

> ...
> I still think it's easier and better to drop it for now. Postponing to 
> the next version is _also_ an acceptable and valid compromise.

It is easier, I'll give you that but I don't see how adding it later would 
be possible without breaking 1.0 clients.
--
Arnaud  Le Hors - Software Standards Architect - IBM Software Group