L2/14-146

Title:  Closing the Donut Hole
Author: Ken Whistler
Date:   June 11, 2014
Status: For consideration by UTC

Problem Statement

During the late phase of preparations for the release of Unicode 7.0,
Asmus Freytag pointed out an anomaly regarding access to the soon-to-be
released version of the standard.

Following our normal process, once the PRI for the beta review for
a version of the Unicode Standard is closed, the PRI link is removed
off the home page, and the PRI is retired to the list of resolved PRIs.
In most cases, this is as expected. But there is an anomaly for the
beta review PRI. This is because during the beta review, the beta
page for the version supplies all the links to the data files under
review, as well as one stop shopping for the proposed updates for
the annexes, and a summary of the major gotchas that implementers
need to watch out for as they are updating.

The net effect of the closing of the PRI and its removal from the
home page is that there is a donut hole where the imminent release
seems to "go dark". Access to all the information about the imminent
release gets harder to find, just at the point when we are trying to
get people interested as we lead up to the release. In the case of
the 7.0 release, the PRI for the beta was closed on May 14, and
the actual release didn't occur until June 16 -- so there was
a period of over 4 weeks when access to the documentation became
more difficult.

In the past, this didn't seem to be too much of a problem. It was
the quiet period when the editors went off and fixed all the documentation
and got ready for the release.

But nowadays, with so many implementers dependent on the releases,
and with the need for access more distributed than it used to be,
it seems unwise for the main Unicode site to "go dark" about a
release between the end of beta review and the actual release.
We could still keep up the appropriate caveats about pages and data
not being final until the actual release is announced, but making
access easier and more continuous with the access we have during
the beta review period would seem a better approach to take.

Suggestion

If folks are in general agreement that we have a problem here, I
would suggest that we ask the editorial committee to adjust the
PRI closure process for the beta reviews for versions of the
Unicode Standard. Instead of just closing the beta review
PRI abruptly, perhaps a two-step process would make more sense,
in which we first stamp the PRI to indicate that the formal
review period is closed, but leave the links and pages active,
perhaps with adjusted notices on them, until very close to
the actual release. Then, just prior to the release, we would do
the final closure of the PRI itself and remove its link to make
way for the release announcement.

Some kind of an approach designed along these lines would give people
a much greater sense of continuity on the site through the beta
review period and leading up to the release. It would eliminate
the donut hole we currently have baked into the process.