RE: Sloppy inference rules

On Thursday, November 01, 2012 5:43 PM, Pat Hayes wrote

> On Nov 1, 2012, at 11:29 AM, Ivan Herman wrote:
>
> > I can see this point. However, my understanding is that a number of
> RDF environments, libraries, etc, have optimized along the lines of
> those restrictions. Changing this, ie, removing the restrictions in RDF
> 1.1, would lead to the necessity of major rewrites of existing systems.
> I would expect that to be a big no-no.
> 
> Yes, I think that is the reason for retaining the restrictions. The
> problem with that line of thinking, however, is that this kind of too-
> expensive-to-change argument only becomes stronger as time goes on, so
> changes will *never* get made. And the counterargument is that allowing
> newer implementations to be more accepting does not invalidate the
> older implementations or make them less able to do anything they can do
> now. It just means that they can no longer claim to be covering all of
> the (now larger) language, which means that the argument is basically
> not even technical, but has to do with corporate advertising and image.

Thanks Pat, that pretty much summarizes what I tried to bring across :-)



--
Markus Lanthaler
@markuslanthaler

Received on Thursday, 1 November 2012 17:02:16 UTC