Waldek Hebisch <hebi...@math.uni.wroc.pl> writes:
| Gabriel Dos Reis wrote:
| >
| > Waldek Hebisch <hebi...@math.uni.wroc.pl> writes:
| >
| > | Note2: I have no idea if/how the construct:
| > |
| > | false == per constantKernel FALSE
| > | true == per constantKernel TRUE
| > |
| > | works in OpenAxiom.
| >
| > Yes, it does.
|
| In FriCAS it would not typecheck. constantKernel is of type %,
| while '%false is of type Symbol.
In OpenAxiom, we have
(1) -> )di op constantKernel
There is one exposed function called constantKernel :
[1] D2 -> Kernel D3 from KernelFunctions2(D2,D3)
if D2 has SETCAT and D3 has SETCAT
In the context where it is used, constantKernel is taken from
`KernelFunctions2(Identifier,%)' because of the import line
import KernelFunctions2(Identifier,%)
Consequently, the expression `constantKernel FALSE' has type `Kernel %'
which is the second branch of the representation domain. Finally the
operator `per' blesses that object of type 'Kernel %' into `%'.
It would have worked the same if I used Symbol instead of Identifier,
but in OpenAxiom we tend to use Identifier when we do not expect
subscripted names (like obviously the case here.)
Any reason why that chain of events isn't happening in FriCAS?
| Even if OpenAxiom compiler
| performs some magic coercions I wonder if result really
| has correct representation.
There is no magic coercions (except that a simple identifier in
a context where Identifier is expected is of type Identifier; otherwise
it is Symbol.) What are the reasons for your wondering?
-- Gaby
------------------------------------------------------------------------------
All the data continuously generated in your IT infrastructure contains a
definitive record of customers, application performance, security
threats, fraudulent activity and more. Splunk takes this data and makes
sense of it. Business sense. IT sense. Common sense.
http://p.sf.net/sfu/splunk-d2d-oct
_______________________________________________
open-axiom-devel mailing list
open-axiom-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open-axiom-devel
