x-samp namespace suggestion

[Mark Taylor]

OK a clarification is a good idea.  How about:
 
    The ``{\tt samp}'' and ``{\tt x-samp}'' form of the same key
    SHOULD NOT be presented in the same map.  If both are presented together,
    the ``{\tt samp}'' form MAY be considered to take precedence,
    though any reasonable behaviour is permitted.

?

On Wed, 28 Sep 2011, Juande Santander Vela wrote:

> Mark, thanks for the clarification.
> 
> I think the “in any case either is likely to be reasonable” makes for a good case for the MAY, but I think it could be spelled out in the sentence, something like:
> 
>   The ``{\tt samp}'' and ``{\tt x-samp}'' form of the same key
>   SHOULD NOT be presented in the same map, but if they are,
>   the ``{\tt samp}'' form MAY be considered to take precedence, 
>   as the final behaviour is likely to be reasonable.
>   
> But now I see why it does not make a difference in behaviour, and it alleviates implementors’ burden.
> 
> El 27/09/2011, a las 19:08, Mark Taylor escribió:
> 
> > Juande,
> > 
> > thanks for the careful reading.
> > 
> > On Tue, 27 Sep 2011, Juande Santander Vela wrote:
> > 
> >> One question: why saying MAY here?
> >> 
> >> The ``{\tt samp}'' and ``{\tt x-samp}'' form of the same key
> >> SHOULD NOT be presented in the same map, but if they are,
> >> the ``{\tt samp}'' form MAY be considered to take precedence.
> >> 
> >> As far as I see it, that MAY should be a SHOULD, to allow for predictability.
> > 
> > My thinking was to reduce the burden on implementors, in cases where
> > the outcome doesn't really matter.  If a map with both keys is presented, 
> > whoever produced it has done something wrong, and the consumer 
> > shouldn't be expected to put much effort into behaving properly;
> > the MAY is just there to give a hint which way to jump if somebody 
> > wants to try hard to do the right thing, but in any case doing either
> > is likely to be reasonable.  If it was SHOULD, then client authors 
> > (with a conscience) might have to spend effort dealing with a case 
> > which is unlikely to occur.
> > 
> >> I cannot foresee an scenario where a client would do different for x-samp.X than for samp.X, outside of system prototyping that should not reach users...
> > 
> > Some kind of bug is the most likely cause; but agreed, it is not
> > expected to happen.
> 
> --
> Juande Santander Vela
> Software Engineer, ALMA Archive Subsystem
> Data Flow Infrastructure Department, Software Development Division
> European Southern Observatory (Germany)
> 
> Anónimo: A veces es necesario guardar silencio para ser escuchado.
> 
> 

--
Mark Taylor   Astronomical Programmer   Physics, Bristol University, UK
m.b.taylor at bris.ac.uk +44-117-928-8776 http://www.star.bris.ac.uk/~mbt/