new TAP-1.1 WD released

Markus Demleitner msdemlei at ari.uni-heidelberg.de
Thu Jul 20 11:06:21 CEST 2017


Dear DAL,

On Thu, Jul 20, 2017 at 12:30:27AM +0100, Mark Taylor wrote:
> here are some comments on the recent TAP-1.1-20170707-WD.

+1 on them from me, and:

> Sec 2.7.2:
>     "... schema, table, and column names, function names, and
>      other ADQL keywords are not case sensitive."
>     I think this neglects the case of delimited identifiers.
>     I will leave it up to Markus or some other SQL ninja to
>     come up with replacement text.

When I read the opening of 2.7.2, I have to say I don't like it too
much any more, mainly because the language makes it seem QUERY is
somehow magically linked to ADQL, and services supporting other
(but similar) languages may want to use something else.

While the "It may also be used..." in the second sentence loosens
things up, I still feel the text could be made more concise.

What about the following wording, replacing the text "The QUERY
parameter ... not case sensitive":

  When, as for instance with ADQL and other SQL-like languages,
  service queries are serialised to character strings in a natural
  way, this query is transported in the QUERY parameter.  Its
  interpretation by the service depends on the value of the LANG
  parameter.  Since ADQL support is mandatory for TAP services, so is
  QUERY.

I frankly would not say anything on the case sensitivity; DALI
already says parameter values are case-sensitive unless specified
otherwise, and I don't think for a database query anyone would do any
sort of case folding in the first place.

A word on case sensitivity *might* be in order for LANG in the
previous subsection, perhaps in the form of

  Note that, by DALI sect. 3.1, LANG values must be preserved
  literally, i.e., no case folding is allowed.

But then I don't think we've had interoperability issues based on
that in 1.0 services and clients, so it would probably be
overcautious to add text like that.

        -- Markus


[1] I'm privately feeling that references to PQL should be removed
entirely from the document; it's not been forthcoming for all these
years, and somehow implying it will, finally, is probably more
confusing than helpful.  It's a private feeling, though, because I'm
not really prepared to work out possible consequences of such a
removal with a standard lawyer's eye.


More information about the dal mailing list