Technical design issues

We can use the [usegalaxy-tools](https://github.com/galaxyproject/usegalaxy-tools/) model for automating the generation and installation (to CVMFS) of reference data.

- For maintenance purposes, since we now have 3 repositories doing very similar things in CI (usegalaxy-tools, idc, and [starforge-recipes](https://github.com/galaxyproject/starforge-recipes)), it might be good to build a common tool to deal with detecting changes and executing a build process
- What do we do when we add a new indexer and need to build indexes for all existing genomes? Either we run this on a small architecture for a long time or a big one for a shorter time. Either way, things like Travis and CircleCI are probably out.
- Should DMs be required and/or preferred? Or should we just use more common tools like snakemake (or just make) w/ Conda.
- ???