PR-TAP-1.1 - Some comments

[Markus Demleitner]

Dear DAL,

On Fri, Sep 08, 2017 at 11:42:49AM +0100, Mark Taylor wrote:
> I always thought that in VOTable (following FITS practice),
> setting the arraysize attribute to "1" has an identical meaning
> to omitting it, i.e. that VOTable makes no distinction between
> a scalar and a single-element array.  However now that I look
> I can't find that confirmed (or denied) in the VOTable specification.

Hmha -- I think the fact that TOPCAT/stilts and astropy, certainly
among the two most widespread consumers of VOTables out there,
disagree on that point is a strong indication that we need to clean
this up.

Anyone in for writing an Erratum?

> On Wed, 19 Jul 2017, Mark Taylor wrote:
> 
> >     In general I think all of the TAP_SCHEMA table columns
> >     defined in sections 4.1-4.4 should be described using the
> >     generic terms either "integer" or "string" (or "character array").
> >     Mandating VOTable datatype and arraysize attributes looks
> >     to me like introducing more problems than it solves.
> 
> I think that listing the VOTable-specific attributes datatype,
> arraysize and xtype in these tables is unnecessarily confusing.

Just to be sure: by "these tables" you do not refer to the TAP_SCHEMA
tables themselves but to the tables in the spec?  If so, then I
agree; let's boild them down to "integer" and "string" (and perhaps
mention that "string" includes what in VOTable would be an array of
unicodeChars; in particular descriptions should be allowed to contain
ødd chåräçtérs

        -- Markus