Datalink Feedback VI: Semantics

[Markus Demleitner]

Dear list,

On Thu, Apr 24, 2014 at 09:29:43AM -0700, Patrick Dowler wrote:
> Q. I did create a table with a column named "property" in postgresql
> and sybase with no issues and didn't find it in the list of SQL
> reserved words I checked... does anyone anticipate issues renaming
> the "semantics" field to "property"?

A quick interaction with the search engine of my choice would suggest
property is a safe column name.  

Still, if everyone could just briefly run

  create temp table autodropped (property real)

on the DB(s) of their choice and shout if there's trouble, I'd say we
can feel even more confident after a few days (incidentally: sqlite
is fine, too).

> Q. Where do we specify the process for updating the vocabulary
> itself? Is that something Semantics WG would have a general policy
> about?

Ah well.  The precedent we have is the UCD addition process, for
which there's a standard of its own, and given that AFAIK nobody has
even tried it I suspect that's not a particularly good precedent.

IMHO, the WG that should determine how to do such things is SDP, and
they should work this out.  In datalink I'd not say more than:

  Service providers are invited to propose additions to the
  vocabulary.  Until the IVOA decides on a formal process for such
  proposals, please direct them to the DAL working group, the contact
  of which can be found on the IVOA page http://ivoa.net; acceptance
  will be decided on by the TCG with single majority.

What I think the SDP should say (and maybe there could be a short,
hackthon-like meeting of the interested parties on this in Madrid?):

(a) After the RFC page becomes obsolete (i.e., with acceptance of the
REC), a <standardname><version>Next page is created on the wiki.

(b) This page collects proposals for addition of features, etc, to
the next version(s) of a standard, proposals for errata, and such, as
well as discussion about them.  This is called the "-Next" page in
the following.

(a) and (b) are, I believe, necessary in general (a Wiki supporting
sub-pages were useful here).

(c) Standards having vocabularies, ontologies, and similar
external semantic resources that should be updateable without
changing the standard itself will have a section on their -Next page
for proposals for each such resource.

(d) The chairs of the working group that issued the standard are
responsible for monitoring these sections and the ensuing discussions
and preparing summaries of them at least one week before a TCG
meeting.  These summaries are distributed to all TCG members with the
expectation that severe disagreements can be clarified before the
actual meeting.

(e) At every TCG meeting, each open proposal is separately discussed.
A proposal is accepted if at least 2/3 of the TCG members present
vote in favour.   If both a chair and a vice-chair of one working group
agree on a veto, this prevents acceptance.

(f) The chairs of the WG that issued the standard summarize the TCG
discussion next to the proposals on the -Next page.


That has the disadvantage that it "constitutionalizes" the TCG, but I
believe it will ensure that vocabulary changes gain enough attention
that obviously suboptimal proposals will usually be improved, so that
may be a small price to pay.

Cheers,

         Markus