NMR language
I propose to create a better language for formal description of NMR experiments than say Bruker or Varian coding language. We will encode NMR experiments using this optimized universal syntax and store them in a web database for subsequent translation (at setup time) into Bruker coding for execution on a given real spectrometer or to Spinach for simulation. Thus, we can store only an blueprint of an experiment removing its immedeate implementational dependances. This will help to optimize individual experiments using theoretcal models.
NMR experiment consists of a basic timing of events diagram known generically as pulse sequence as such and a calibration convention connecting abstract entities as a 90deg pulse to shaped radio frequency burst of given strength (in db units) and time duration specific to a given spectrometer as well as some auxiliraly files like composite pulse decoupling sequences inserted as wholistic locked units within the basic time diagram.
A certain high level (generalized) representaion of experiments ideally should be human oriented with dynamically accessible context information (like spin state at this time or pulse). In one of my projects (submitted to Bruker, but rejected) we proposed even to generate NMR experiments directly from chemical structure, e.g. just mark spins to be correlated and get working experiment. At this time I was a member of Informatics Institute of Uni Zurich, people their liked this kind of global ideas and were ready to try to implement.
We should list requirements for such a language first, of course.
Well, after some thinking, I propose the radical solution — I will write a compiler that would take pulse sequence description and generate the effective Liouvillian stack. It seems there’s no other way than the one musicians are taking — they have a basic time grid and all events are snapped to it and discretized on it. The grid will be adaptive and will have a form of a vector of timings plus a stack of matrices. Run that through expv and we’re done.
