Transf WD Where are the axis

[David Berry]

On Fri, 3 Apr 2020 at 13:14, Laurent MICHEL
<laurent.michel at astro.unistra.fr> wrote:
>
> David,
>
> My feeling is, like Gerad said, that the mappings should be more
> explicit in the model. I missed that point on March 10.
>
>
> Le 02/04/2020 à 16:55, David Berry a écrit :
> > I agree maybe we've been a bit too free and easy with the use of the
> > word "axis" in the document. I many cases we should probably replace
> > "axis" with "mapping input" or "mapping output".
>
> Would be better but likely not enough (IMO).
>
>  >
> The definition of
> > which physical quantities are associated with each input and output is
> > provided by the TransformSet.
> Do you mean TCompound?

No, I really do mean TransformSet.  Are you sure you're looking at the
new draft?

https://volute.g-vo.org/svn/trunk/projects/dm/STC/Trans/doc/WD-Trans-1.0.pdf

> I'm may be confused with the scope of the model.
> In FITS-WCS there is no explicit binding between e.g. image pixels and
> WCS keywords.
> WCS just says that this image has been generated by applying that
> transformation. If the scope of TRANSF is to do the same in VOTables, it
> is fine.
>  From another hamd, if the ambition of Transf was to be able to model
> workflows (sequence of transformations), we would have to be able to say
> what kind of input data are expected by a given operation and to which
> data are produced. If we assumed that all transformed things are
> matrices, just specifying the input/output dimensions could do the job
> does it?

That is  the purpose of the TransformSet - to define what physical
quantities are expected as the inputs and outputs of the Mappings.

David