On 01/10/2014 05:56 PM, W Brown wrote:
> On Jan 10, 2014, at 4:16 PM, John Spicer wrote:
>
>> On Jan 10, 2014, at 3:09 PM, "Nelson, Clark" <clark.nelson_at_[hidden]> wrote:
>>
>>>> I was wondering if it would be possible to have a github project
>>>> for the SD-6 document from this group?
>>> That's a very good question, which I have been planning to discuss in
>>> Issaquah.
>>>
>>> In a lot of ways, we're in uncharted territory. Most standing documents are
>>> strictly administrative. This is the first one I know about that actually
>>> has technical content. Because of that, there needs to be a process in place
>>> to approve changes to it. So we don't want to make it *too* easy to modify.
>>>
>>> On the other hand, I don't think we want to take a committee vote every time
>>> we want to revise it. The membership of WG21 endorsed the approach, so as
>>> long as we stay consistent with that, we should generally be OK. I don't
>>> know, maybe changes should be approved by consensus of SG10 participants.
>>>
>>> The fact that it's now hosted on isocpp.org is also relevant. From a
>>> technological perspective, I don't know whether that helps us or hurts us.
>>> It's basically a wiki with an approved list of editors, and every member of
>>> the committee is expected to become an approved editor. (Of course there
>>> will be other approved editors as well.) I don't know whether it also has
>>> revision control features.
>>>
>>> That is substantially what I have been thinking about the question. Thanks
>>> for raising it on the reflector.
>>>
>>> From a committee process point of view, I think a consensus of SG10 participants is probably a good way to go.
>> An addition to that could be asking that people proposing new features get input from SG10 about whether a new macro is needed, and if so what its name would be. That could then be included in the paper voted on by the committee so people would be exposed to the name and given an opportunity to comment.
>>
>>> From a technical point of view, it might be good to have it under version control (in github or elsewhere) and then updates would be copied to the isocpp web site.
> a) FWIW, SG10 already "[hopes] that every paper will include its own recommendations concerning feature-test macro names" [N3745]. Perhaps whenever Clark in future announces or reminds WG21 of upcoming mailing deadlines, he might also remind authors about this SG10 request.
>
> b) It seems to me that SG10 might in future routinely schedule a between-meetings teleconference to review the most recently approved WG21 motions for their possible impact on our document. At the following WG21 meeting, SG10 can then report on the existence of an updated version of SD-6.
>
> c) I would fairly strongly prefer that changes to SD-6 occur only as a result of SG10 consensus, and that such approved changes be committed only by the SG10 chair or other SG10-designated individual. I would be very leery of allowing changes to any standing document by an arbitrary WG21 participant, no matter how well-motivated.
I support this. I was just asking for a process where the larger group
could post patches to the list, discuss them and refine them. Then
someone would either send up a pull request to github or the last patch
on the list would be blessed. The final patch is taken by the person
with write permission could have the final say and make the final
merge. I wasn't proposing a free-for-all ;-)
>
> -- WEB
> _______________________________________________
> Features mailing list
> Features_at_[hidden]
> http://www.open-std.org/mailman/listinfo/features
>