Re: summary on options for JP-4 Comment about the semantics of property paths

On 11/02/12 16:55, Andy Seaborne wrote:
> == Short version
> I can live with either option.
> I prefer option 2 when done properly but as things stand, I have to say
> that I can't see option 2 + quick 3LC as viable.
> I've implemented, experimentally, a form of option 2. It's available to
> all.

 >>Just leaving things as they are, and pointing to a future WG is 
another Option. Should we call it Option3?

My understanding of option 1 was that it was to leave things as they 
are, with editorial text to point that some property paths are expensive 
and that SELECT DISTINCT can be used to control this, but to claim there 
was a set of query transformations that are equivalent to the possible 
semantics in the paper.  This is what Axel is calling option 3 now [1].

Given Axel's clarification: I want to restate my current position:

I can live with option 2 or option 3.  Option 1 is tolerable but it not 
currently sufficient detailed to know whether it is viable.

Option 1 would be OK if there were a set of transformations but to work 
them out it seems we might as well do option 2.  The fact that option 1 
might avoid a LC seems less important if it needs time to formally 
define and then check the transformations and their consequences.



Received on Sunday, 12 February 2012 10:14:55 UTC