Pondering over strategies for evaluating BLIPKIT vs Pellet for use with NMRShift data

As a subproject, I'm now focusing on doing some comparison between reasoning in Prolog and Pellet, using example data from NMRShiftDB.org. Below I'm documenting my planning, + any new findings while I'm digging in to this subpeoject.

Objective of subproject

To evaluate Prolog/Blipkit based querying against NMR Shift data, compared with doing the same in Pellet. The evaluation should focus on time of execution, and expressiveness/ease of implementation (possibly compared as lines of code, etc.).

Things to decide

  • What format to accept for an incoming spectrum, for comparison against e.g. other spectra?
    • Probably best to start with accepting "array of arrays" as that is the output from rdf.sparqlRemote().
  • How to preprocess / store the data for use in Prolog?
    • What methods are there available in Blipkit for this?
  • How to preprocess / store the data for use in Pellet?