W3C home > Mailing lists > Public > wai-xtech@w3.org > April 2012

Re: Versioning ARIA editor's drafts

From: Benjamin Hawkes-Lewis <bhawkeslewis@googlemail.com>
Date: Sat, 21 Apr 2012 03:50:49 +0100
Message-ID: <CAEhSh3c5XKvrgng4M96zA_M_HxCyE44g7tqerkbDkn3cwb5Fww@mail.gmail.com>
To: Michael Cooper <cooper@w3.org>
Cc: PFWG Public Comments <public-pfwg-comments@w3.org>, W3C WAI-XTECH <wai-xtech@w3.org>
On Fri, Apr 20, 2012 at 9:23 PM, Michael Cooper <cooper@w3.org> wrote:
> One thing we *can* do is put the CVS version identifier in the document,
> near the top. Although it isn't possible to view a specific historical
> version online, the version identifier at least allows a commenter to be
> specific about which version they were reviewing. I can also send a copy of
> the relevant version on demand. As a short term solution, would this help?

Citing a revision number would be a marginal improvement over citing a date.

HTML WG publishes a revision number for HTML5 Editor's Drafts.

But if there's no public access to the revision, reviewers are better
off continuing to make public archives with tools like WebCite than
involving W3C staff in copying revisions around.

> One other thing we sometimes do is create a dated editors' draft, to a URL
> like http://www.w3.org/WAI/PF/aria/YYYYMMDD/. Normally we do this only for
> "milestone" drafts where we want review of a stable version, but aren't
> publishing a TR version. I'm reluctant to generate dated editors' drafts
> every time we check in a change to the source, because the number of copies
> would proliferate and become hard to manage. But I'll ask, is this something
> you are looking for? If so, how important would that be to you?

This would enable permalinks to a specific version without using an
external archiving system, so that would work.

> Longer term, I've been considering putting the ARIA documents into the W3C
> Mercurial repository. That repository *does* provide public access to
> historical versions of the document, though I find it can be a pain to
> obtain the proper URL.

Could you surface the URL in the published draft?

> My thought had been just to post the source versions
> there, not the generated editors' drafts; the source versions would be less
> useful to reviewers.

Again, this would enable permalinks to a specific version and changes
without using an external archiving system, so that would be an
improvement. This is effectively the system HTML WG/WHATWG use.

A big advantage of a public repository is changes could be watched
using source control tools rather than by comparison of HTML pages, so
I'd be in favour of this over versioned URLs for Editor's Drafts, if
you were going to pick one.

> But we could potentially put generated editors' drafts
> into that repository if it will be helpful to reviewers. If we do this, will
> it meet what your looking for?

I think this would only be worth it if the repository browser rendered
these drafts as HTML with working links, so that one could provide
deep permalinks into a working document.

--
Benjamin Hawkes-Lewis
Received on Saturday, 21 April 2012 02:51:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 13:16:16 GMT