On Sat, 16 Jan 2016, Allin Cottrell wrote:
On Sat, 16 Jan 2016, Sven Schreiber wrote:
>
> What about scalar versus series? "x = 5" can also mean either. Is this
> also covered by the changes?
[the changes that soften the scalar/matrix distinction for
automatically types variables]
No. I believe it has always been clear that if you want a series of constant
value 5 you must say
series x = 5
or
x = 5 * const
and I don't see any reason to change that.
Maybe that's too brusque. I seem to recall that you (Sven) posted a
comment on series versus scalars not so long ago, and though I don't
remember the details, it may have raised an issue that is not
answered above -- in which case please come back on this.
My overall view is: if the user hasn't specified a type for a new
variable, the type assigned should be the "natural" type based on
the right-hand side expression. That's not always totally clear, but
it seems to me that "x = 5" naturally yields a scalar, without
ambiguity.
Allin