Skip to content

Setup for Windows #7

@matuskalas

Description

@matuskalas

I went through the https://github.com/NBISweden/workshop-python/blob/gh-pages/ht19/preliminary.md setup guidelines, concentrating on the implications for Windows users (as I'm one of those, despite of being a big Open Everything enthusiast :D )

It might of course be a bit late now to do any updates to the instructions, something for you the mainteners to consider, but in any case here follows a list of possible issues I found, that I worry may cause some difficulties for the tech non-savvy novices. So good to be aware in advance.

In order of appearance in preliminary.md:

  • Python: There are 4 installation options for Windows at https://docs.python.org/3.7/using/windows.html#installation-steps, each of them problematic, unstable, or just messy in a different way. That's why we have conda and Anaconda these days! Therefore, it might be something to consider to ask the participants to just install Anaconda, at least for Windows user, or perhaps also MacOS, or even Linux, too. That's also how The Carpentries do it by default (https://carpentries.github.io/workshop-template/#python). The only real downside of installing Anaconda is that it takes upto 6GB space and a lot of time.

  • Dependencies: With Anaconda installed on Windows, pyreadline, pandas, and matplotlib are installed already and can be imported immediately. No pip install needed. (And it's also discouraged in the conda philosophy to install packages into the base environment.)

  • Text editor: On Windows, Notepad++ is a very nice editor for programming and handling textual data files. Again, also recommended by The Carpentries (https://carpentries.github.io/workshop-template/#editor).

  • Test: On Windows with Anaconda|Miniconda, it's the best to use the Anaconda Prompt terminal. Then no messing with environment variables is needed. python3 does normally not work on Windows, and in the cmd terminal it may then open Microsoft Store to install the unstable Python 3.7 from there :D Use just python instead.

  • Jupyter: Jupyter Notebook is included in the Anaconda install. (Plus the "Running the notebook" link doesn't seem to point anywhere useful and is not on the github.io page. On Windows with Anaconda, one can just click to open the Jupyter Notebook shortcut, or otherwise just type jupyter notebook in the Anaconda Prompt).

If not now, these might be something to consider for the future.

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions