Motion 42 (not 41): Decoration system, revised text

[John Pryce <j.d.pryce@xxxxxxxxxx>]

P1788

First, apologies that for some time I thought this was motion 41, and put this in my email headers.

As a present to you all before my Christmas break, I send a revision of the set-based Decoration System text under discussion for Motion 42. It is labeled Draft 6.2. Recall that the current vote is on the general principle; however, I think this text is approaching its final form as standard text.

 Main changes are as follows.
(1) It is now split into a flavor-independent introduction which is part of Chapter 1 (currently Clause §6) and the set-based part (currently §8.8). Namely, the flavor-independent part (I hope) of §8.8.1 has moved to form §6, leaving §8.8.1 as a short stub. 

(2) A number of changes have been made as a result of discussion. Chair: I hope it's OK for these to be friendly amendments.

Attachment 20121221DecoSystemCirculatedA.pdf holds §6, and also §5 on Flavors, since they need to be read together. 20121221DecoSystemCirculatedB.pdf holds the revised §8.8.

For now, please vote on §6 and §8.8 together as specifying the set-based decoration system. A future motion will decide whether §6 is OK as a flavor-independent specification of decorations. BUT comments on that will be most valued.

The main changes (2):
 - "Permitted combinations" is rewritten. There are now 4 subclauses 
  8.8.3. Ill-formedintervals
  8.8.4. Permitted combinations
  8.8.5. Initial decoration
  8.8.6. Decorations and arithmetic operations
where before there were two, I think this is more logical and I hope 8.8.5 meets Vincent Lefevre's objections.

 - The com decoration is more fully described and integrated into the text.

There's still quite a lot of detail to put in, e.g.
 - specification of decorated constructors,
 - operations to extract interval & decoration parts, 
 - behaviour of comparisons & other operations that don't return an interval
...

If anyone feels that they would vote Yes, except for the fact that certain items are still missing, now is the time to itemize these key items. And, please, specify the changes that would lead you to vote Yes. The sooner we get such comments the quicker an agreed text will be reached.

One other thing. I revised §5 Flavors to define the notion of a "loose common evaluation", which is basically a common evaluation done in finite precision. I think this lets one tidy up a lot of verbiage, but I may have missed something, so please look at this, and how it is used, critically.

Merry Christmas to all

John Pryce

Attachment: 20121221DecoSystemCirculatedA.pdf
Description: Adobe PDF document

Attachment: 20121221DecoSystemCirculatedB.pdf
Description: Adobe PDF document