W3C home > Mailing lists > Public > public-w3process@w3.org > October 2014

Fixing the draft in TR space is continuously out dated problem

From: Harry Halpin <hhalpin@w3.org>
Date: Mon, 27 Oct 2014 05:18:23 +0100
Message-ID: <544DC78F.9090700@w3.org>
To: public-w3process@w3.org
Everyone,

Inspired by yet another really fun time trying to get a WD published, I
propose that we get rid of the step of publishing separate WDs from the
Editors Draft in /TR space entirely.

I think the publishing of WDs before CR is a technical procedure that
would have limited or no formal effect on W3C process and considerably
negative effects on the standardization process insofar as it confuses
developers/users and wastes the time of editors/Team. The formal impact
is limited as I believe patent commits are made to the charter and the
call for Exclusions happens at FPWD and then patents commits again at
PR->Rec. Although I may be wrong, in between charter, FPWD, and CR no
patent commits or exclusions generally are made.

Currently, the Working Draft publication process is at some date we take
an arbitrary snapshots of the Editor's Draft (currently almost always
maintained in Github rather than W3C.org space), do a week-long e-mail
dance with pubrules, link checker and the Webmaster, and then publish a
Working Draft in TR space - at which time this WD is usually *out of
date* by the time it appears on W3.org.

This by necessity leads practical WGs to emphasize a "living spec"
procedure where we encourage people to look at Editors Draft and ignore
whatever WD is at W3.org/TR space. No one really tracks the old
"hearbeat" requirement anymore, as the action is in the bug tracker and
Editor's draft.

Further, it is made difficult to publish WDs in W3.org because of small
issues with our somewhat clunky pubrules process and the fact that link
checker is often excluded by robots.txt (for example, from fragid links
to github) and the Webmaster basically acts like a robot that just
checks to see Link Checker and Pubrules is fulfilled but cannot edit the
HTML. It seems like if the Webmaster just acts like an automated checker
for this clunky process, there's no reason not to replace the Webmaster
with a fully automatic process for WD publication, as that would save
both the Editor and Team time and the Webmaster could use their skills
for something more important than running Link Checker, like fixing the
rest of the W3C toolchain.

So, a simple proposal to fix this situation is that:

1) After FPWD when a /TR URI is given, a simple automated snapshot of
the Editor's Draft from Github (or wherever else it's maintained) is
just sent automatically to the right w3.org/TR space. Thus, no more
confusion between WD and Editor's Draft, because there would be *no
difference*. At FPWD, a Call for Exclusions is sent.

2) Publishing WDs should no longer require any interaction with the
Webmaster. The Team Contact or Editor should just be able to run
pubrules, Link Checker, or whatever else they like and use their own
judgment if they think something needs to be fixed - in which case, that
fix should be reflected back into the Editor's version on Github.

3) The Webmaster and the full publication process is only required at
the transition out of WD into CR, PR, and Rec (with usual calls for
Exclusions).

Does anyone see any utility in keeping publishing WDs separate from
Editors Drafts?

Apologies in advance if this has been discussed before.

     cheers,
       harry
Received on Monday, 27 October 2014 04:18:26 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:35:12 UTC