Drafting a short-term roadmap

[oesteban]

With the new momentum we are gaining with the contributions from @esavary and @jhlegarreta, I think it would be great to look forward and devise a more updated roadmap.

I've created a new milestone 0.1.16 and assigned there some issues I believe we could address first.

I think we should start with #152, just to take a snapshot of today and flex the release workflow.

I've left a few issues out, but feel free to add them, or create new ones.

cc/ @arokem @effigies

[oesteban]

Also, setting a concrete goal could help move this forward. What about the following goal for 0.1.16 -- a release that:

• can be integrated with MRIQC (i.e., no packaging issues, uploaded to pypi)
• can replace 3dVolreg, currently used there, to estimate head motion
• does the estimation faster than 3dVolreg (which should not be that hard because it is single-threaded in MRIQC, so it takes long). And to be faster,
  • it leverages parallelization of registration (through ANTs),
  • a coarse-to-fine approach (ENH: Data manipulations to facilitate modeling and registration: smoothing & subsampling #116),
  • and brain masks.