Rough time estimates for developer time for PACE

Go back: Horace framework redesign ideas

---

Rough time estimate for effort required for each task / work-package:

• Implement the distributed computing mex module for the current Horace Matlab code (AB estimates 3 dev-months)
• Refactor cut_sqw and sqw_eval to use the distributed computing modules (either pure Matlab prototype or new mex one; AB estimates 2 dev-month)
• Refactor the current Horace data classes to better conform to OOP standards (AB estimates 1 dev-month)
• Implement generic projections (AB estimates 5 dev-months)
• Implement refactoring the sqw binary file format to use hdf5 (AB estimates 1 dev-month)
• Implement a Python interface to a compiled Matlab library version of Horace (MDL estimates 1 dev-month)
• Rewrite the core of Horace in Python (?18-54 dev-months?), consisting of:
  • Rewriting the core classes, sqw and dnd, including sigvar and arithmetic operations (3-6 dev-months?)
  • Rewriting the graphical plotting engine, including the 3D sliceomatic plots (3-6 dev-months?)
  • Rewriting the multifit interface for fitting (3-6 dev-months?)
  • Rewriting / extending TobyFit (3-12 dev-months?)
  • Rewriting the interface for the current C++ mex files (3-12 dev-months?)
  • Rewriting the GUI (3-12 dev-months?)

---

Go back: Horace framework redesign ideas