Erratum uploaded

[Markus Demleitner]

Dear Colleagues,

On Tue, Nov 04, 2014 at 05:28:10PM +0530, francoise.genova at astro.unistra.fr wrote:
> How do we make sure that an erratum does not have adverse effects?
> I understand that for you it may not be so important because you
> consider that errata come and go. I think that we need to have at

Just to prevent a discussion about a position I don't hold: Of course
errata need to be throroughly reviewed.  Posting them on a wiki page
certainly won't make them that.

The way I'd like things to be is that on -Next pages, there'd be a
special area (say, "Endorsed") that may only be written to after a
review process (I'd say consent from both WG chairs and TCG review
are minimum with some additional formal criteria, cf. below), and
everything outside of this is considered draft or discussion.

As to extra formal criteria, I am sure that in the review process
implementations are the most critical thing, or maybe a close second
after validators.  At least the two-implementation rule should be
adopted; I think I'd even say "Maintainers of actively developed
implementations and validators of the standard have veto power."  I
believe any changes for which strict consensus might become a problem
shouldn't be an erratum in the first place.

But of course text review is important as well (well, usually the
most thorough text review comes with implementation), as shown in the
current example, where Mark Taylor spotted an ugly bug in the text
(Volute rev. 2764).

So -- I'm not arguing against review and consensus building at all,
on the contrary.  I just believe errata are something else then first
class citizens of the IVOA document repository in many respects and
should be treated as such.  And conversely, treating them as normal
documents requires constant bending of both the document and the doc
repo, which just doesn't feel right.

Cheers,

        Markus