- From: Francois Daoust <fd@w3.org>
- Date: Fri, 21 Aug 2020 09:04:36 +0000
- To: "Corentin Wallez" <cwallez@google.com>
- Cc: public-gpu <public-gpu@w3.org>
Oh, I didn't mean to scare anyone! It is absolutely great to have active individual contributors from the community and we're certainly not going to interrupt that with deterrent "Join the WG or leave!" messages. Please keep on contributing! I cannot think of written guidelines. This situation is not yet a common one in W3C: in most cases where there is both a CG and a WG (Immersive Web, Second Screen, Audio, etc.), work items either belong to the CG or to the WG. WebAssembly is the only real other case where work items are shared between a CG and a WG. I'm not aware of major issues having arisen in the WebAssembly case. We're still learning along the way. In the end, it depends on what "active" and "individual" mean. My "join the WG" encouragement is intended at people working for companies that have a vested interest in the API (and possible patents!), for instance implementers and hardware manufacturers. It seems to me that the majority of people who regularly attend GPU for the Web group calls and contribute actual normative contents to the WebGPU specs fit in that category. As far as I can tell, the underlying companies are also W3C members, and it should be relatively straightforward for them to join the WG. We need IP commitments from these companies to help guarantee that the specs can be published under a royalty free license in the end. The sooner, the better. On top of IP commitments, it seems important to make sure that the WG and CG represent the same GPU community, and that decisions taken in one group reflect the intents of the other. For instance, that would help avoid running into situations where there exist frictions between participants that crystallize into an objection to include a particular feature being raised at the WG level but not at the CG level, preventing publication of the spec as is, and creating more friction. Francois. ------ Original message ------ From: "Corentin Wallez" <cwallez@google.com> To: "Francois Daoust" <fd@w3.org> Cc: "public-gpu" <public-gpu@w3.org> Date: 20/08/2020 18:34:02 >Hey Francois, > >Thanks for all the information! > >One question about asking active contributors to join the WG, are there >written guidelines at when to do this? We might get some active >individual contributors from the community and joining the WG might be >a bit scary because of all the IP commitment (when it's very unlikely >it will actually affect them). > >Cheers, > >Corentin > >On Thu, Aug 20, 2020 at 4:14 PM Francois Daoust <fd@w3.org> wrote: >>Hi WebGPU enthusiasts, >> >>As relayed by Xueyuan, the GPU for the Web Working Group (WG) has now >>officially been created. As discussed during last Community Group (CG) >>call, the idea is to make the operation of the WG as transparent as >>possible so that you can simply continue to work on the WebGPU and >>WebGPU Shading Language specifications without having to worry too >>much >>about process implications. >> >>In particular: >> >>- The public-gpu@w3.org and internal-gpu@w3.org mailing-lists are now >>associated with both the CG and the WG. Whenever a person joins the CG >>or the WG, they will be subscribed to these mailing-lists. >> >>- The idea is to also share the repositories under the gpuweb >>organization on GitHub. I will prepare pull requests to add `w3c.json` >>files to relevant ones (see doc at [1]) so that W3C tools can process >>them. I will also work with chairs to add gpuweb/gpuweb to the >>repository manager (see doc at [2]) to enable automatic IPR checks on >>pull requests. The checks will flag pull requests that come from >>non-WG >>participants (because that is where the IPR regime is the most >>restrictive and the spec eventually has to be published on the W3C >>Recommendation track). That does not mean that non-WG participants >>cannot send pull requests, just that you will have to pay attention on >>whether you deem their contributions acceptable for publication under >>the W3C Patent Policy. One practical consequence is that I will ask >>main >>contributors to associate their W3C account in our systems with their >>GitHub account so that the bot does not flag false positives. If not >>already done, you can do that through your "Connected Accounts" page >>[3]. >> >>On top of my head, a few changes that may be worth noting: >> >>1. Horizontal reviews and resolution of issues raised during >>horizontal >>reviews are now a must to progress on the W3C Recommendation track. >>Given the specs at hand, some reviews will probably not trigger much >>feedback (Accessibility? Internationalization?), while others may >>raise >>more thorny issues (privacy/security, TAG). >>2. The decisions to publish and progress the specs on the >>Recommendation >>track are in the hands of the WG. >>3. I will serve as W3C staff contact for the WG. The expectation is >>that >>I will only bother you for process-related issues, and sometimes >>editorial issues on the specs. You will likely not see me much during >>technical discussions though (and that's a good thing, I'm no expert >>in >>GPUs ;)). I am not planning to attend weekly calls at this stage for >>instance, unless some process question needs to be discussed. >> >>Essentially, CG/WG boundaries can remain pretty transparent provided >>that active CG participants are also WG participants, which seems >>doable, on paper at least. In short, expect me to chase you down if >>you're actively involved in discussions and have not joined the WG yet >>;) >> >>The home page of the WG is at: >>https://www.w3.org/2020/gpu/ >> >>(W3C Members can join the group from that page. Note that your >>Advisory >>Committee representative is the only one who can nominate people to a >>group). >> >>Let me know if you have questions! >> >>Thanks, >>Francois. >> >>[1] https://w3c.github.io/w3c.json.html >>[2] https://w3c.github.io/repo-management.html >>[3] https://www.w3.org/users/myprofile/connectedaccounts >> >>
Received on Friday, 21 August 2020 09:04:43 UTC