Ralf Hemmecke wrote:
>
> I don't like Union(SomeDomain, "failed") very much.
>
> 1) Since Union actually requires domains as arguments, there must be
> some special compiler code dealing with that.
Yes, there is special case code, but it is probably more general
than you think: each string is regarded as being a type having
itself as unique value. If context requires String as a type,
then this unique type is silently coerced to String. Consider:
(2) -> "ralf"::"ralf"
(2) ralf
Type: ralf
(5) -> List("ralf")
(5) List(ralf)
Type: Type
(8) -> "ralf" has Type
(8) true
Type: Boolean
(9) -> "ralf" has SetCategory
>> System error:
The index 4 is too large.
The last result shows that this "string as a type" mechanism is not
fully implemented.
ATM I am not sure what to do about this. On one hand it allows nicer
syntax (without a lot of boilerplate to declare trivial types).
On the other hand it introduces irregularity into types system
and it is not clear how much effort is needed for complete implementation.
> In the long run, I would
> like to either have a special domain Failed, which only has one element
> called failed and which can be used as Union(D, Failed) or to have a
> domain like Partial (see below).
>
> 2) I've tried to write a Partial domain, but with some problems.
>
> ---rhxBEGIN partial.spad
> )abbrev domain PARTIAL Partial
> Partial(V: Type): with
> if V has CoercibleTo OutputForm then CoercibleTo OutputForm
> inject: V -> %
> retract: % -> V
> failed: () -> %
> failed?: % -> Boolean
> == add
> Rep := Record(val: V)
> if V has CoercibleTo OutputForm then
> coerce(x: %): OutputForm ==
> failed? x => "failed"::Symbol::OutputForm
> x.val :: OutputForm
> inject(v: V): % == construct v
> retract(x: %): V == x.val
> failed: % == NIL$Lisp
> failed?(x: %): Boolean == EQ(x, failed)$Lisp
> ---rhxEND partial.spad
>
> (1) -> Integer has CoercibleTo OutputForm
>
> (1) true
> Type: Boolean
>
> (2) -> Partial Integer has CoercibleTo OutputForm
>
> (2) false
> Type: Boolean
>
> Any idea why I don't get true for (2)?
>
As I wrote in proivate mail you currently need Type as part of
category expression.
> 3) I actually would like something like in Aldor, i.e. that I can specify
>
> failed: %
>
> in the categorial part. But that doesn't work. Any chance that it can
> work in some near future?
>
But what the above do. Consider the following implementation:
Rep := Integer
n := 0
failed == (n := n + 1; n)
Is it allowed by Aldor? Related to this is question if runtime
interface is functional (this is each time you use 'failed' getter
function is called to fetch its value).
--
Waldek Hebisch
[email protected]
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/fricas-devel?hl=en.
