SIAP and encoding of POS input parameter

Alberto Micol alberto.micol at eso.org
Wed Jan 21 11:26:01 PST 2009


Stating the obvious...

1.- The Standards is the standard, one cannot dispute it [xml schema 
datatypes]

2.- Programmers are programmers (you can dispute them, but it won't help)
      Sometimes a careful computer scientist,
      sometimes an astronomer,
      sometimes the daughter of an astronomer...
      and we all know that the daughter of the astronomer will not 
assume what her parent
      gives for granted ("I do not need to read what a 'double' is, I 
know it!")

3.- VO -which will probably never have a full set of validators in 
place- MUST work.

4.- Even Doug had to dig into various sections of the W3C standard to 
understand
    that the canonical representation might differ from the lexical one. (?)
    Could you imagine the typical astronomer's reaction after digging 
into those W3C and IVOA
    "insane" documents?


All this brings me obviously to the conclusion that we must be very 
explicit.

- ADQL is explicit, and uses a clear BNF
  (if I get it right, it is actually the canonical representation 
mentioned by Doug).

- Doug is pragmatic: 

    If one is writing a client app the safest thing to do is omit the
    leading "+".  If one is writing a service the safest thing to do is
    accept it. 
    (btw, true for the input query params, for the output params the
    opposite applies)

I think it is important to repeat both these things IN EVERY document 
-not just into a note
that might end up being not read even by the careful astronomer's daughter.

Alberto



More information about the dal mailing list