Re: Two Specs and Test Suite

On 2/28/2012 1:18 PM, Anne van Kesteren wrote:
> On Mon, 27 Feb 2012 21:09:58 +0100, Jeni Tennison <jeni@jenitennison.com>
> wrote:
>> Therefore, I propose that we do both. Further, I'm happy to edit the
>> declarative-style specification unless there's someone else who wants to
>> do it. Are there any objections to that as a way forward?
>
> Sounds great to me. If we have the resources to research both, why not. I'm
> very much with Mohamed that all the focus on process (in other threads)
> seems somewhat out of place in a CG. I'm curious as to what your draft will
> look like :-)

Great.

I've been among those who have advocated trying a more declarative approach 
along with Anne's more imperative formulation, but in the interest of full 
disclosure...

In general, declarative or functional mappings are more easily applied to 
parsing input that is correct (and reject input that isn't) than it is for 
expressing fixups of messy input. So, I'll be really glad if we can find a 
more declarative way to do XML-ER, and I think we'll learn a lot by trying, 
but I won't by surprised if we fail. After all, there are certain things 
for which you need the power of a Turing-complete language. Probably it 
will come down to what sorts of fixups we decide are appropriate.

Noah

Received on Wednesday, 29 February 2012 00:02:02 UTC