Title: The Standard Report

Author: Bob Schmidt

Date: 02 May 2020 18:07:36 +01:00 or Sat, 02 May 2020 18:07:36 +01:00

Summary: Guy Davidson reports on the changes that have been brought about as a result of the CoViD-19 pandemic.

Body: 

According to the schedule we were following in February, the next face-to-face meeting was due to take place in Bulgaria at the beginning of June, in the coastal resort of Varna. As you can imagine, several of us were looking forward to a little extra time in the summer sun by the sea on either side of the meeting.

There was considerable concern, of course, as the coronavirus pandemic unfolded before us. We debated on the mailing lists about how we should proceed, should we cancel immediately, wait it out, take a view three months ahead or some mixture of these approaches.

Cancelling a meeting is not a trivial matter. They are expensive to host, requiring 10 conference rooms to accommodate over 200 people for a week. Attendance isn’t cheap either: flying out to these locations can be an expensive proposition for self-funded attendees, so buying flights early is an important saving. Several committee members had therefore already bought flights and were reluctant to see their investment set aside by cancellation.

Ultimately though, amid conference cancellations and rumours of impending lockdowns, the decision was made for us when ISO announced that all face-to-face meetings were cancelled until the end of June. This date has since been extended to the end of July. There is no sign of this edict being lifted, and nobody is assuming that the November meeting in New York is going to take place.

We still have a standard to deliver in 2023. The ‘train model’ of standard delivery, where features ship in a standard if they are ready or hop on the next train to ship in the next standard if they aren’t, rather relies on a solid three-year cycle. The motivation behind this approach is to ensure regular updates to the standard, rather than a repeat of the ’00s where nothing was released except for the Technical Report in 2003.

This means that we still need to carry out the work of the committee but not in face-to-face meetings. Just as we are all getting used to teleconferences over Zoom, Teams, Hangouts, Webex and all the other video meeting platforms, so must the committee, and that nettle has been well and truly grasped. A new schedule of meetings has been devised, with weekly meetings of Library Evolution and Evolution. The schedule oscillates a little to ensure that it doesn’t exclude people with other regular meetings, although it does run into the problem of having an international community, which means time zones.

Library Evolution meetings are scheduled for alternating Mondays at 15:00 UTC and Tuesdays at 17:00 UTC, while Evolution meetings are scheduled for alternating Thursdays at 17:00 UTC and Wednesdays at 15:00 UTC. This works out at early morning for the US West Coast, lunchtime for the US East Coast, teatime for the UK and dinnertime for Europe (or thereabouts).

The meetings run for 90 minutes and, at the time of writing, we’ve just finished the second week of this regime. It seems to be working out well. Initially I was rather aghast at the idea of remote meetings for the standard: it is easier to block out a week three times a year than it is to block out two ninety-minute meetings twice a week. I was concerned that experts would be absent from critical meetings and important decisions would be made without sufficient oversight. However, the purpose of these meetings is not to ram proposals through the pipeline, but to smooth over as many rough edges as possible, of as many proposals as possible, so that when we do finally meet face-to-face, there will be less need to spend committee time on individual papers.

The hope is that we can continue with our schedule of works for C++23 which you can see in P0592, available at https://wg21.link/P0592. In summary, we are prioritising the following for inclusion in C++23:

• Library support for coroutines
• A modular standard library
• Executors
• Networking

You will see that this list contains no language features. I’m quite pleased about this: C++20 was a huge upgrade to the language, and I think the user community needs time for the wave of upgrades to finish crashing and to recede a little.

We also expect to make progress on:

• Reflection
• Pattern matching
• Contracts

Reflection and contracts are handled in their own study groups, SG7 and SG21, rather than in EWG, but we expect to see papers emerge from those groups for EWG to consider, although not necessarily for inclusion in C++23.

Of course, this is NOT to the exclusion of all else. If you are interested in, say, ooh, I don’t know, the linear algebra proposal, then that will continue to be discussed. There are several active subgroups still reviewing proposals: numerics, HMI, game-dev and low-latency, text and machine learning, to name a few, still have a queue of proposals to consider. They are likely to emit something for the consideration of LEWG or EWG: it will be up to the chairs of those groups to arrange proposals according to the priority listed above.

It’s certainly not business as usual for the committee. However, we have a job to do, to continue improving C++ and shipping new standards every three years. We have adapted to new ways of working, increasing the use of teleconferencing and holding more frequent, shorter meetings to meet this goal. The committee process has never been more accessible: if you have ever wanted to participate and been thwarted by the cost of travel and accommodation, now is the time to test the waters by joining the meetings and experiencing the process.

If you need more details about how to do so, please contact me or the editor. We would be happy to point you in the right direction.

Please stay safe and healthy!

Guy Davidson is the Principal Coding Manager at Creative Assembly. He has been writing games for about 75% of his life now. He is climbing through his piano grades and teaching Tai Chi while parenting in Hove and engaging in local political activism.

Notes: 

More fields may be available via dynamicdata ..