outline.md

title	author	progress
Astronomy and “eResearch”: My Journey	James Tocknell	false

But few of us have been taught the necessary skills…

---

I’m going to cover what I think everyone needs to know

. . .

And how I learnt these skills.

---

---

Use version control for everything

• Scripts, code, papers, everything
• github/bitbucket/gitlab have free private repos for students/academics

---

Use a package manager

conda, spack, pip, apt, homebrew, chocolatey

Package your code

• Python: Python Packaging User Guide
• Fortran/C/C++: autotools, cmake
• See https://swcarpentry.github.io/managing-research-software-projects/

Manage your data

Automate

• Make
• cookiecutter
• Testing—see https://katyhuff.github.io/python-testing/

---

So how to learn more?

---

Courses

Books

• Effective Computation in Physics
• Python Cookbook
• The Hacker's Guide to Python

Other people/Meetups

• Coffee and Code @ MQ
• Hacky Hour @ UQ
• Other unis?
• ResBaz

Internet

• Stackexchange (stackoverflow, tex, unix, and more)
• The Hitchhiker’s Guide to Python
• Python Module of the Week
• Scipy Lectures
• PyVideo
• Blogs: Python Planet
• Project documentation

---

xkcd 1205

---

Questions?

See https://aragilar.github.io/MtStromloXmasSeminar2017Talk/ for slides

Other resources

• autotools book autotools advice
• unix advice and more
• Why not csh
• reproducable research
• how to report bugs
• git for ages 4 and up
• git for computer scientists

Why not conda?

• See KERN paper for why they didn’t use conda
• Not everything in conda (hence KERN)
• Not as powerful as other systems (e.g. Spack, nix)
• Does not work with some python tools
• Try conda, but be aware of alternatives