Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: Four vs. five decorations



Nate, Dominique and all

On 30 Jun 2011, at 15:05, Nate Hayes wrote:
>> I've yet to see a counter-example that shows how *any* use of intersection
>> and union with FTDIA gives valid results, except by accident or coincidence.
>> 
>> So my conclusion is the theory as presented is invalid.
> 
> BTW, just so we are clear:
> 
> I'm assuming by "stated conditions" you are referring to the conditions outlined in Motion 26, notably section 5.8.6 of v3.02 that John just circulated on Jun 20, 2011.

I find your scheme is angled toward particular union/intersection scenarios also and will give false results when applied to other scenarios.

My position at present is that union and intersection should ALWAYS return a bare interval. It is up to the programmer to decorate it appropriately for the application.

My rationale is that ALL the various examples so far shown build to the conclusion that union/intersection are not free-standing operations in the way that, say, an interval extension of a point function is. Their decoration logic is too contingent on the environment in which they are used, to be fixed in this standard. If we do fix it, these operations will become illustrations of the proverb "If all you have is a hammer, everything looks like a nail".

An alternative is to have different versions of union & intersection whose decoration behaviour is correct in different contexts. But for KISS reasons I say no thanks to that idea.

I seem to disagree with Arnold AND Nate on this, but there it is.

John