what are our use cases?

[jovo]

store use cases

required functions (ndstore)

1. find data
2. get metadata
3. store images
4. store annotations (paint+metadata)

   additional functions

5. read volumes for ML (ndstore)
6. write volumes off microscope (ndingest)
7. read tiles (ndtilecache)
8. write annotations (ndblaze)
9. microservices
   1. downsample/upsample
   2. compute histograms

      explore use cases (all ndviz)

10. visualize image tiles & stats & metadata
11. visualize volumetric annotations & metadata

    analyze use cases

12. volume recon (not done yet)
13. 2D histogram correct (misha2d)
14. 3D histogram correct (misha3d)
15. object detect (ndparse)

    model use cases

16. anything bobby did (ndpaper:cs1)
17. 3D point process model of synapse locations (ndpaper:cs2)

[wrgr]

Ok, after lots of reflection and cleaning up the draft I think it's most clear to structure as follows:

• backend centric stuff in store (this is all db stuff, storage, novel read/write algs, etc.)
• explore is ok as is
• analyze contains all the stuff it has now, but with the addition of restful endpoints and ndio - I think it's much cleaner and also most useful for explaining the new paradigm (e.g., this is how people use stuff for science, image processing, etc.)
• model is ok as is. Adding text to explain case studies are modeling in current draft

Let's chat this morning? slightly differently in that I would move the user facing stuff to analyze. Which is to say: store is all the "backend" stuff that users would expect on a backend .