Re: Use of dated drafts (Re: Expiration for certificate management)

On Thu, May 28, 2015 at 11:07 PM, Harald Alvestrand <harald@alvestrand.no>
wrote:

> Replying to the PS, so changing the subject.....
>
> On 05/29/2015 02:16 AM, Martin Thomson wrote:
> >
> > p.s., Apparently the certificate management changes I proposed have
> > been merged, but I can't see them online.  The editor's copy is
> > outdated.  Other specs are now editing directly on gh-pages, which
> > means that the editor's copy is actually current, as opposed to some
> > arbitrarily old copy.  I just looked at the process that is being used
> > here and it's quite an unusual process.  I can only ask: why?  It's
> > got to be much more work this way.
> >
> You can inspect the editors' working copy by pulling the main branch.
>
> The purpose of using dated versions is to make sure that all listed
> editors have actually had a chance to review what goes out in their
> name; at least one listed editor has said that this is an important
> consideration.
>

Hmm... I don't really care about this, but on JSEP we deal with this by
requiring all editors to agree before PRs are merged, so that the
head revision always has rough editor consensus.

-Ekr


> The separation between "working copy" and "published draft" should also
> give a chance for making sure the "published draft" is not in an
> inconsistent transition state in the midst of a change, but I'm not sure
> how much we're actually succeeding in that.
>
> Note: Before github took over the world, dated editors' drafts seems to
> have been the norm.
>
>
> --
> Surveillance is pervasive. Go Dark.
>
>
>
>