hi josé, trevor & all
is the main issue the overhead created by setting up the development environment on a host of different machines?
I recently spent a couple of years teaching a bunch of computer music classes at a community college without
any proper facilities and was faced with a similar problem. My own experience as a student was always in a college
with a dedicated computer room with a bunch of desktops that had pre-installed software and it meant that we could
focus on music related computer stuff, rather than spending half the semester trying to figure out how to set up a particular
environment.
I'm not sure what sort of solutions yourself and Trevor have already discussed, but I was wondering if it might
be possible to set up a cloud based solution to your problem? Setting up a container for abjad development
would be an interesting little exercise, and it shouldn't be a huge stretch to get this running on a server where
users could access using ssh. You could use a fresh container image per cohort or whatever, so it would be
easy enough from an admin perspective for you to manage the server.
The fact that abjad outputs pdfs to its root directory by default is also pretty handy, because it means that any
sort of a simple socketserver running in this directory would make the pdfs directly accessibly through the browser.
Also, I don't know if part of the discussion was whether or not maintaining ipython support was a bit of a crick in the neck.
I imagine that possibly replacing this with a single docker image and then encouraging users to pursue a modular workflow
might be less work?
Let me know what y'all think!
regards,
Adam