HaloO,
I wrote:
John M. Dlugosz wrote:
> I wrote a complete treatment of Inf support.
> Please take a look at "24.26 Infinite" on pages 116-119, and
> "3.11.3 Infinities" on pages 26-27.
I have a lot to say to that. Please give me time.
I find your treatment of Inf too Num centric. E.g. already the
built-in Complex should have an unsigned infinity. Or one that
has an argument. This argument might also be useful for the
complex zero. In both cases you have something like a lenthless
direction.
In general the Undef and Inf types should be parametric on the
underlying type of concrete objects. Undef is sort of not yet
in the type and Inf is outside the type "on the other end" so
to speak. So I think Undef and Inf span the type.
E.g. sqrt(2) might return a special Inf that can be lazily
stringified to an arbitrary long sequence of digits of sqrt(2).
Assignement to a Num collapses the Infinity towards the nearest
Num. IOW, there are Aleph1 infinities of Num between 1 and 2.
That is what *defines* Num in contrast to Rat which is Aleph0
infinity as a whole. But perhaps we should call this Aleph1
complete type Real and not use it ;)
Also you missed transfinite ordinals which can be very useful.
I'll be offline for two weeks, TSa.
--
"The unavoidable price of reliability is simplicity" -- C.A.R. Hoare
"Simplicity does not precede complexity, but follows it." -- A.J. Perlis
1 + 2 + 3 + 4 + ... = -1/12 -- Srinivasa Ramanujan
