Reorganizing this repo as a continuing resourse for computing training for incoming KIPAC folks at all levels.

[eacharles]

This issue is just to track suggests and changes that we might make to reorganize this repo as a continuing resource for training new kipac folks, including incoming summer students, grad students, post-docs & staff, who obviously all have different requirements.

[sidneymau]

Reproducing my comments in Slack here.

Some brief observations following the last two days:

• A lot of the students already have a decent level of computer literacy and are comfortable with python, notebooks, some basic scientific computing APIs, etc. (sessions weren’t as helpful for these students)
• Many of the students have resources and mentoring from their groups (duplicated effort)
• A small number of students are really starting from scratch (these students probably need/needed more individual help than what these sessions could provide)

Suggestions going forward

• Send out a poll in advance so we know roughly the level of computer literacy/experience as well as operating systems coming in (calibrate to the students and make sure we’re prepared for how to help them)
• Rather than a general computing setup BootCamp, we run office hours for the first week or two to help those out who are starting from scratch or running into bugs (help the students who need it and use the scaffolding from the groups rather than try a “one size fits all” solution)
• Start with sessions that are related to tools more specific to astro research than just Python. A number of students expressed an interest in learning statistics and how to do statistics in Python, for instance—i.e., we can start from import scipy.stats as st rather than print("Hello world!"). Something like Adam’s StatisticsCookbook, for instance

Another thought would be to keep the format of running general sessions in the first week or two, but to have the first week just be getting new students up to speed, and then in the second week jump into something like statistical analyses with Python, for instance.

Some concrete to-dos:

1. Define the scope of what we want the students (at different levels) to get out of the BootCamp. E.g., we want undergraduate summer students to be comfortable running Python code and debugging problems they run into; we want graduate students to know how to write software packages with proper testing and distribution via pip, and so forth.
2. Move to using a website (this repository has a live gh-pages branch deployed here: https://kipac.github.io/BootCamp/) for organizing information and linking to notebooks. I think the GitHub directory format is a bit confusing, especially for new students. This is related to gh-pages has out of date (2018) schedule/registration #39.
3. Prune the repository of older tutorials. These will still be available through Git commit history, so this is a reversible change and one that would be very helpful in terms of making this task more approachable.
4. Rewrite new tutorials in-line with the scope defined above. We should strongly consider avoiding duplication of effort by leveraging tutorials available from common resources (e.g., NumPy's own tutorials) and focus our efforts on things more specific or unique to computing at KIPAC.

I am happy to contribute to all of 1–4 above, but I think (personally) starting with 2. might be a good idea, if there is consensus that this is a good thing to do, as it might help us think about how we want to organize/focus/highlight different parts of these resources.