[J3] Paper 22-120r4 available

[Van Snyder]

On Wed, 2022-03-09 at 18:32 +0100, Magne Haveraaen via J3 wrote:
>    My opinion is that we should start with a small, efficient and
> effective template feature. Then we can extend it later on. Trying to
> put everything into it now will make it much more complex to specify,
> implement and use. And we will create lots of horrible corner cases
> that become impossible to solve/retract later on (confer experience
> from C++ evolution).

Ada generic package didn't appear to have the "growing pains" that C++
templates had. I don't remember any significant revisions between 1983
and 2012. 

>    What the generics subgroup is proposing and that Tom has worded is
> close to a small, efficient and effective template feature. Expanding
> that, like adding TBP, seems to incur a significant amount of
> specification work, more unforeseen issues and a heavier
> implementation burden. Extending templates at a later stage, when we
> have more Fortran experience, will be a much safer, stepwise
> evolutionary approach which contributes fewer pitfalls.

I don't think type-bound procedures are a problem, if they're described
correctly: The binding in the definition for a template parameter that
is a type specifies the required interface for the binding in the
actual parameter, but not the name of the specific procedure. The
syntax and constraints on the PROCEDURE statement within a type
definition would need adjustment to allow it to refer to an interface
body, not a procedure.

Generic bindings might be a problem, if you want to solve the problem
Magne mentioned about < actually being >, or + actually being min.

> 
> Magne
> 
> 
> > 8. mar. 2022 kl. 14:37 skrev Clune, Thomas L. (GSFC-6101) via J3
> > <j3 at mailman.j3-fortran.org>:
> > 
> > I have modified the paper to split the initial straw vote into 3
> > separate straw votes.  Allowing nuanced differences in the
> > treatment of type-bound procedures, data components, and
> > kind/length parameters.
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.j3-fortran.org/pipermail/j3/attachments/20220309/3cbb21e1/attachment.htm>