[J3] (SC22WG5.6218) Re: (SC22WG5.6217) All face to face INCITS or US-hosted ISO meetings cancelled for the rest of 2020
Ondřej Čertík
ondrej at certik.us
Thu May 21 11:01:22 EDT 2020
Hi Steve and others,
On Thu, May 21, 2020, at 7:52 AM, Steve Lionel via J3 wrote:
> On Thu, May 21, 2020 at 1:56 AM Bryce Adelstein Lelbach aka wash via J3
> <j3 at mailman.j3-fortran.org> wrote:
> >
> > INCITS has cancelled all face to face meetings of INCITS committees and US hosted ISO meetings for the rest of the year. Just wanted to give y'all a heads up as I didn't see a post to your mailing lists about this yet.
>
> This came out just yesterday afternoon, my time. I immediately wrote to
> Dan Nagle and the J3 subgroup heads to get some initial thoughts on how
> J3 and WG5 should handle this. I haven't heard from Dan yet, but did
> get some ideas from others.
>
> From the WG5 side, which is where I could make a decision, my
> inclination is to just cancel the WG5 meeting for 2020. WG5 doesn't
> have any real business at this time - we've already agreed on the 202X
> work list, aren't discussing 202Y proposals yet, and don't have a
> corrigendum to discuss. So even a virtual WG5 meeting seems less useful
> at this time. Our 2021 meeting is scheduled for late June in
> Manchester, UK, and I very much hope the world will be in a place by
> then where we can do that.
I would suggest to keep the (virtual) meeting, it can be shorter if needed. If it is held, does it have to be 5 days long? I didn't see this specified in the ISO rules, but maybe I missed it.
If it is true that "WG5 doesn't have any real business at this time", then I suggest we use the effort to either help the 202X standard out of the way sooner, or start tackling proposals for 202Y. In particular, the most requested difficult problems seem to be:
* generic programming
* exceptions
* GPU support
An alternative solution is to just held the J3 (virtual) meeting, but invite all WG5 members to participate.
> When WG5 and J3 meet together, the WG5 members are very helpful in
> contributing to J3's work on technical papers, but that can still
> happen. There's already been a lot of discussion on the J3 email list
> about providing J3 an online space for discussing/developing the
> technical work. The idea I have is to set up a discussion board on the
> J3 website for J3 (and WG5) members to do the work that usually happens
> in subgroup meetings. My hope is that we can transition to doing more
> of this work between meetings than we have been in the recent past, but
> this can't be a complete substitute for the personal interactions we
> get when we meet physically - I hope we can get back to that next year.
Steve and I discussed this privately and it seems we we agreed on the general framework how this can work. There are two distinct issues:
1. Moving the work of the Committee itself online. This must follow the ISO rules and it must restrict the discussion to committee members. There are multiple ways to do that, from a private GitHub or GitLab repository, to some other private forum such as Discourse or phpBB. Or even just an email on a private mailinglist. This is what Steve's paragraph above suggests.
2. Involving the community. This is what the "incubator repository" (https://github.com/j3-fortran/fortran_proposals) is for. I do advocate the committee members to *optionally* use it and to discuss as much as they are able and really to embrace it as a tool to develop proposals together with the community. This is distinct from 1. and is not supposed to replace 1, but rather complement it --- to feed into 1.
We need both 1. and 2.
We might disagree on the extent to which the community should be involved, in other words how much work will be done using the 1. versus 2. approach, but we made huge progress on figuring this out and agreeing that we need both 1. and 2., and I suggest we simply start doing it and figure out and adjust the details as we go.
Ondrej
More information about the J3
mailing list