| 1 |
% --------------------------------------------------------------------- % |
| 2 |
% DIRECTORY: CPO % |
| 3 |
% % |
| 4 |
% DESCRIPTION: basic theory of CPO's and fixed-points % |
| 5 |
% % |
| 6 |
% AUTHOR: Albert J Camilleri % |
| 7 |
% % |
| 8 |
% ADDRESS: Hewlet Packard Laboratories % |
| 9 |
% Filton road % |
| 10 |
% Stoke Gifford % |
| 11 |
% Bristol, BS12 6QZ % |
| 12 |
% England. % |
| 13 |
% % |
| 14 |
% email: [email protected] % |
| 15 |
% % |
| 16 |
% DATE: 90.06.01 % |
| 17 |
% --------------------------------------------------------------------- % |
| 18 |
|
| 19 |
I have found the [contributed] theory very useful for defining non primitive |
| 20 |
recursive operators. I use it to define the recursive operators in my latest |
| 21 |
model for CSP and I intend to install it in the library in some future release |
| 22 |
of HOL. In the mean time, however, it might come in handy to other HOL users. |
| 23 |
|
| 24 |
The [contributed] HOL theory is simply intended to lead to the fixed point |
| 25 |
theorem which will allow the definitions of recursive operators. It is not |
| 26 |
intended as an exhaustive theory on domains, etc. Users may need to prove |
| 27 |
further theorems as required, but I think that [it] gives a good start for |
| 28 |
defining recursions. For example, in CSP an ordering on processes (<= say) is |
| 29 |
defined and proved to be a CPO, a `bottom' process is defined and proved to be |
| 30 |
so, and then for all functions f:process->process, recursion can be defined as |
| 31 |
(CONTINUOUS f) ==> (MU f = FIX f $<=) and the definition of FIX can be |
| 32 |
expanded to the value of the least upper bound. |
| 33 |
|
| 34 |
Hope it comes in useful. |
| 35 |
|
| 36 |
Albert |