Re: RIF Core, and how much is PRD allowed to diverge from BLD [Was: Re: [PRD] Issues to resolve before publication]

Gary Hallmark wrote:
> To maximize rule interchange between production rule engines and logic 
> rule engines, clearly Core should be "as big as possible". 

Yep. Why not equate PRD with BLD? That would indeed mawimize rule 
interchange between BLD-compatible rule engines. Why do we need PRD, 
after all?

Of course, that would make useful interchange between PR engines more 
difficult, and they might choose not to use RIF for that purpose... But 
who cares?

It is not like we also wanted to maximize utility for PR engines, of 
course: that would make it a multi-criteria optimization problem, and 
those are notoriously difficult...

> We can, 
> should, and must decide that now.  I don't even know why I have to keep 
> arguing this point. 

Did you ever argue it? You repeat it ad nauseam (even menacing to shout 
it, sometimes :-), and you keep ignoring my arguments in support of a 
different view (some of which seem pretty sensible to me, though)...

(But maybe that is only my impression. Maybe, as seen from your side, I 
am the one not listening. Is it the case?)

> The bias to keep BLD and PRD aligned with a large 
> common core should be so high that the burden of proof is on you to show 
> why NAU  should not be in Core.  You have provided no such proof.

I have explained why I believe that maximizing the interoperability 
between PRD and BLD did not mean either blindly aligning PRD on BLD 
without taking PR specifics into account, nor talking BLD-ese to the PR 
crowd.

Re the "burden of the proof", a tit for a tat: as you well know, no 
proof holds against a dogma. As you have erected blind alignment of PRD 
on BLD as a dogma, the burden is on you to convert me...

:-)

> I am much more interested in making PRD useful for exchange with a 
> variety of rule engines than I am in tailoring PRD to any one or two 
> vendors products.

One benefit of tayloring PRD to the mainstream PR engines, including the 
one or two vendors you have in mind, is that it augments the chances 
that they will adopt and deploy PRD.

Indeed, the balance is between tayloring PRD to the mainstream PR 
engines and making it interoperable enough with BLD.

But tayloring PRD to the mainstream PR engines come first, I believe, 
because, on the one hand, the interoperability has benefits only if PRD 
and BLD are adopted and deployed; and because, on the other hand, it 
will be easier to increase the interoperability (in future versions or 
further dialects) once it has proven its benefits, where, again, 
adoption and deployement is a prerequisite.

Let me put it otherwise: what is the variety of rule engines you have in 
mind, besides the mainstream PR engines I have in mind?

Cheers,

Christian

Received on Tuesday, 1 July 2008 12:25:54 UTC