[Omake] Re: [CUCS Bug 545] "or" and "and" take each argument to be a _sequence_ of values.

[Jason Hickey]

Bugzilla is down again:/

In any case, it really isn't cuteness that is important.  Whatever we do 
should:
    - be simple
    - be some model of a Boolean algebra, where [[and]]=/\ and [[or]]=\/

I suppose we could say that [[and]]=f for some other (f : bool list list 
-> bool), but it seems nice to me that f should be a (generalized) model 
of /\.

The usual axioms like the de Morgan's laws, units, associative and 
distributive laws, etc. should hold.  You might argue that some laws are 
not important, but I don't see much sense in choosing "false" as the 
identity for conjunction.

Jason

Bugzilla - Please Use the Web to Reply wrote:
> Reply at: http://cvs.cs.cornell.edu:12000/bugzilla/show_bug.cgi?id=545
> 
> 
> 
> 
> ------- Additional Comments From nogin at cs.caltech.edu  2006-06-26 14:17 Eastern -------
> As I said, your option 2 (which is the same as my option 2) is my least favorite
> option (aqmong those that I consider acceptable).
> 
> Your option "1" is indeed "cute", but it is simply not what many people would
> expect, so for me it falls into the category "not the way to go, no matter how
> nice it looks".
> 
> There are at least two more options that I would consider acceptable. Consider
> (for sequences of boolean values), strong-and(e1 e2 ... en) = n>0 /\ e1 /\ e2 /\
> ... /\ en. Then the options (unsing the numbering I gave in comment #6) are:
> 
> 1) [[$(and e1, ..., en)]] is strong-and([[e1]])/\.../\strong-and([[en]])
> 
> 3) [[$(and e1, ..., en)]] is strog-and(strong-and([[e1]]) ... strong-and([[en]])
> 
> And, BTW, there is no need to spend 8 lines on option 2:
> 
> 2) [[$(and e1, ..., en)]] is strong-and([[e1]]), when n = 1
>                              jason-and([[e1]] ... [[en]]) when n != 1
> 
> 


-- 
Jason Hickey                  http://www.cs.caltech.edu/~jyh
Caltech Computer Science      Tel: 626-395-6568 FAX: 626-792-4257