Discussion
Loading...

Post

  • About
  • Code of conduct
  • Privacy
  • Users
  • Instances
  • About Bonfire
Adam Steer
@adamsteer@mapstodon.space  ·  activity timestamp last week

All the people in my feed loudly complaining about people using Jupyter notebooks, I'm curious:

What's your proposed fix / replacement for doing a document which has words and runnable code snippets?

How are you proposing to do better?

...that isn't "use different coding language" or "just these seventeen steps to cloud deploy the discombobulator"

If there's something, I'm sure people would love to know!

#python #JupyterNotebooks

  • Copy link
  • Flag this post
  • Block
Konrad Hinsen
@khinsen@scholar.social replied  ·  activity timestamp 2 days ago

@adamsteer Here is my proposed fix:

https://hyperdoc.khinsen.net/

However, that's a fix for only one use case of notebooks: scientific publishing. I see the main issue with notebooks not in the technology, but in the widespread enthusiasm to use them (or advocate for using them) for just about everything.

Lab scientists don't publish their lab notebooks, they write papers based on data from their lab notebooks. Personal notes and papers require different media. Also for computations.

  • Copy link
  • Flag this comment
  • Block
Adam Steer
@adamsteer@mapstodon.space replied  ·  activity timestamp last week

Here's an example of what I see as a core use case for Jupyter notebooks. The concept is kind of a "step through runnable paper". Aimed at people who are research specialists not software developers.

Idk... Open, reproducible, shareable science 🤷. It isn't finished - contract ended so more #lostscience

https://gitlab.com/npolar/aen/sea-ice/-/blob/main/jupyter-notebooks/AeN-ice-density.ipynb?ref_type=heads

  • Copy link
  • Flag this comment
  • Block
Adam Steer
@adamsteer@mapstodon.space replied  ·  activity timestamp last week

...and I have *seen* some things in research code that notebooks are an absolutely vast improvement on. Mainly looking at you, senior researchers who gronk out Matlab with two letter variables and no comments ever, and then gaslight the fuck out of people for not being smart enough to reverse engineer code you don't even understand anymore and mostly didn't in the first place.

So from that foundation, Jupyter notebooks have a lot to like. Far fewer excuses for not explaining ones self.

  • Copy link
  • Flag this comment
  • Block
Adam Steer
@adamsteer@mapstodon.space replied  ·  activity timestamp last week

My own solution, what I propose, to improve Jupyter notebooks?

It isn't technical. It is human factors work - nurturing the discipline, the time, the working style needed to use the tool effectively. Less about the latest/greatest, more about how and why.

...which would also apply if a person chose a different tool (R, matlab, whatever)

Anyone want to hire me for that?I'd be great at, say, running a cold regions research group and offering mentoring / courses on how to do all this...

  • Copy link
  • Flag this comment
  • Block
Log in

bonfire.cafe

A space for Bonfire maintainers and contributors to communicate

bonfire.cafe: About · Code of conduct · Privacy · Users · Instances
Bonfire social · 1.0.0 no JS en
Automatic federation enabled
  • Explore
  • About
  • Members
  • Code of Conduct
Home
Login