[J3] work on F202X at #217?

[Bill Long]

Hi Anton,

This is my take on the outcome of the Berkeley meeting and J3’s job for meetings 217 (Oct 2018) and 218  (Feb 2019):

Our primary task is to provide more specific proposals for WG5 to consider for the F202X content at the Tokyo meeting in 2019.  It would be desirable to have a feature set pretty much settled at that meeting.  Features need to have a good reason why we need this capability - i.e. Dan’s “use cases”.  I think it is also helpful for J3 to provide technical input on issues like implementation difficulty, whether the feature is conflicting with the existing standard, and integration difficulty for adding it to the standard.  The old justification of “we can’t think of a reason to not do this” is no longer adequate.  

Input to the process seems to come from 3 places:

1) The parts of 18-156 discussed at Berkeley:

     The generics/templates idea got good support (Y:14, N:4, U:3), but we should try to narrow in on which direction to go.  Use cases will be particularly important. 

     The exceptions proposal got thinner support (Y:7, N:5, U:9).  I think this one needs more discussion.

     The “undecided” parts of 18-156:  I think that += , *=  (Y:3, N:10, U:5),  Constructors (Y:1, N:8, U:12), and allowing specifications in arbitrary blocks (Y:4  N:8, U:7) got poor support.

     The “undecided” part of 18-156 that did got a positive vote - allowing arrays of a derived type with coarray components (Y:8, N:5, U:7) needs to be fleshed out in more detail.

2) The parts of 18-156 that were NOT discussed at Berkeley:

      A fairly long list that all got positive votes from J3.  These are basically the “US proposal”. 

3) Input from other countries.  These are due 6 months after the Berkeley meeting, so maybe not considered until m218 next February.  Hopefully some will arrive soon enough for m217.



Secondarily, assuming that F2018 is published before the February meeting, some Interp activity on the new standard may require attention at m218.



 
> On Jul 20, 2018, at 6:03 AM, Anton Shterenlikht via J3 <j3 at mailman.j3-fortran.org> wrote:
> 
> I'd like to know what work
> is going to happen on F202X at
> meeting 217 in LV?
> 
> In JUN, WG5 asked J3 to work
> on generics, exceptions and
> arrays of allocatables with 
> coarray components.
> The medium/small features mentioned in 18-156
> were not discussed.
> 
> Will J3 start drafting work on
> generics and exceptions?
> Will this be done via papers?
> Will the work be split into sub-committees?

I asume any output from the meetings will be in papers, and that subgroups will address technical issues relevant to their subgroups. 

> 
> Realistically what progress
> can be expected at 217 towards F202X?

I hope quite a bit.  I’d hate to have to do it all at 218 in February.

Cheers,
Bill

> 
> Thanks
> 
> Anton

Bill Long                                                                       longb at cray.com
Principal Engineer, Fortran Technical Support &   voice:  651-605-9024
Bioinformatics Software Development                      fax:  651-605-9143
Cray Inc./ 2131 Lindau Lane/  Suite 1000/  Bloomington, MN  55425