[Psidev-ms-dev] Some additional 'Unknown instrument' CV parameter thoughts

[Lennart M. <len...@eb...>]

Dear PSI-MS Enthousiasts,


After Tuesdays (5 Feb 2008) phone conference, I've given the addition of 
an 'unknown instrument' term some further thought, since it made me feel 
ever so slightly uncomfortable to add this 'escape hatch' to the CV.

Upon consulting with Luisa Montecchi about this, she suggested a 
potentially viable alternative: why not simply use the top-level CV term 
instead? So whenever the actual instrument model is not known, we would 
annotate as 'MS:1000031 -- instrument model'.

I see two benefits in this approach:

  1) no need to include a term called 'unknown instrument' in the CV, 
which to me at least feels a bit awkward;
  2) the validator can cope easily with allowing/disallowing this parent 
term while allowing all children regardless, whereas allowing all but X 
children of a term (as we would have to do in the case of adding an 
'unknown instrument' child term to 'instrument model') is possible, but 
ugly.


There is also a downside to this approach -- which I will also try to 
subvert into an upside :)

  the knowledge that the top-level term is to be used for 'unknown 
instrument' will be documented in the spec doc, but it is likely that 
few people will read this document in earnest. So the ability to flag an 
instrument as 'unknown' will be an 'expert feature' only. The upside of 
this is that it provides a ready incentive to convertor implementers and 
  instrument manufacturers to go the very short extra distance to 
accommodate (interactive) user input of the instrument, or addition of 
the new instrument model to the CV, respectively. Interestingly, 
enabling/encouraging both of these solutions were actually raised as 
important points during the phone conference.


What are your opinions on this proposal?


Cheers,

lnnrt.