NXtransformations validation problems

[mkoennecke]

The NXDL for NXtransformations is IMHO wrong:

1. The required content of NXtransformations is a depends_on NX_CHAR field. This is not there.
2. Then there is a template for an anonymous field with the CIF attributes in there. IMHO, this is
   purely documentation. While, most fields involved in transformation should be in NXtransformations,
   the depends_on chain can also point out of the NXtransformations group.

I want this to be seen and OK'ed before I fix this

[zjttoefs]

1. The NXtranformations group is a container for transformations. As such it has no defined origin of it's own. All transformations contained need to be valid individually and thus have a depends_on. So I have no idea what a depends_on in the NXtransformations would mean.
2. Nothing is said about where the dependency chain points. There are no limits there. The specifications means to say that this group contains any number of transformations (ideally at least one though) without requiring certain naming. If that can be better expressed, that would be okay with me.

I guess we can close this after the next telco.

[zjttoefs]

The missing depends_on is correct, we agreed on in the telco.

The anonymous fields still pose a problem in the validation. There is a suggestion by Peter Chang for a similar problem elsewhere that was accepted by Pete Jemian. I will dig that out in a minute.

[zjttoefs]

It was actually exactly this case. We even have an issue already (now closed): #272
The name tag for what is thought to be a viable solution was accepted in pull request #411
Closing here (duplicate now).