To: J3                                                     J3/25-158
From: Brad Richardson and Generics
Subject: Relitigating the Characters Selected to Bracket Instantiation
         and Deferred Argument Lists
Date: 2025-June-25

Reference:
  24-125r5
  https://tinyurl.com/44tje3rp

1. Introduction
=========================================================

In paper 24-125r5, it was selected to use parenthesis to bracket
the list of deferred and instantiation arguments in templates
and instantiations. Subgroup still feels that curly braces
would be better, and this paper intends to re-litigate the
issue.

2. Rationale
=========================================================

The decision was highly contentious, and there was no
strong preference determined by the results of the straw
votes.

The inclusion of the circumflex (^) character in front
of the instantiation arguments for inline instantiation
is both superfluous and utilizes a previously unused
character. It is also not particularly ascetically pleasing.

One of the primary arguments against utilizing the curly braces
was that they are difficult to distinguish from parenthesis.
However, parenthesis are even more difficult to distinguish.

The broader Fortran user community seems to prefer the use
of curly braces, as can be observed based on discussions
on the Fortran-Lang.org Discourse forums, starting
at the following link: https://tinyurl.com/44tje3rp.

Additionally, the author of the only prototype currently in existence
that demonstrates the feature would prefer curly braces.

3. Straw Vote on Deferred/Instantiate Argument Delimiters
=========================================================

What symbol(s) should be used to delimit deferred and
instantiation argument lists?

A. Parenthesis with ^ added for inline instantiation: ^()
B. Curly Braces: {}
C. Undecided

Examples:

A.

   subroutine swap(T)(a, b)
   call swap^(T)(a, b)
   call P1^(P2^(x))(a, b)

B.

   subroutine swap{T}(a, b)
   call swap{T}(a, b)
   call P1{P2{x}}(a, b)