This sounds like an excellent idea, and something that we could definitely use in teaching here at Berkeley!
Are you focussing on simple in the UI view, or simple elsewhere…
eg this repo branch (which works in MyBinder) sets up a Jupyter environment using Jupytext that lets you author markdown files (.md
) and run them, via Jupytext, as a notebook, but the save format is just .md
(no ipynb). To create a new notebook-from-md, create an .md
file and then click on in the link to it in the Jupyter notebook home page to view / edit / run it as a notebook.
Thanks for sharing feedback all. Our goal of the project is to uncover barriers that can make JupyterLab difficult to use. We hope to use these insights to learn more about how people use JupyterLab. Here are some questions that we have:
Have you used Classic Notebook? How does it compare to JupyterLab?
What type of work are you doing when using JupyterLab(education, industry, etc)?
Are there other software tools that are important for you to complete your work?
I think @Eric_Van_Dusen @yuvipanda and @willingc could be interested in these questions from their experiences in education w notebooks
Hi - Yes - we use classic notebook for teaching a lot. The idea in some ways is specifically to not give a first-time user a view that looks like and IDE but a view that looks like a written page - so that the familiarity and lack of options will focus them on some simple learning outcomes. In some ways our view is about giving novice users exposure to datascience in the least intimidating way - not even mentioning much background just dropping them into a notebook and have them clicking and learning one command at a time.
To all who have given feedback, thanks so much . I have been toying with the idea of Progressive Disclosure as a good lens to solve this type of problem. I am also looking into features that are currently underserved.
One thing to think of for inspiration might be the Adobe suite of tools. I like how they have different layouts you can customize and then make then one click away, as well as a set of out-of-the-box layouts for new users -> more advanced users.
Thanks for all your feedback y’all. We have taken the direction of making this a “Single Document Mode”. Would love to hear some thoughts.
Also, this project has begun a conversation of what it looks like to install packages with a UI. Here is a gif of what it looks like. We are working at getting a binder link working.
oh wow! I think it looks great! Really nice job
Do you have a demo somewhere for people to try? Maybe a Binder?
We’ll get that to you soon! Thanks for waiting.
I’d love to have a look at it once it is published somewhere!
This looks really nice @javag97!
I missed this thread when it was posted a couple of months ago. Does it correspond to this repo? GitHub - jupytercalpoly/jupyterlab-clarity-mode: A document-focused, decluttered mode of JupyterLab that uses activity-based design
In the meantime there has been ongoing work on a project called JupyterLab Classic.
It’s alternative JupyterLab distribution with the look and feel of the Classic Notebook (a remix of existing JupyterLab plugins):
It’s currently being developed in GitHub - jtpio/jupyterlab-classic: The next-gen old-school notebook UI 📓, and can easily be installed with pip
and conda
/ mamba
:
pip install jupyterlab-classic
mamba install -c conda-forge jupyterlab-classic
Do you plan to continue working on the focus mode project? It would be interesting to compare the approaches, and maybe share components or development efforts.
Thanks!
FYI, note that the date is 2019, so more like 1.5 years ago…
Ah that must be why!
to avoid confusion for others like me who would like to give this a try:
It seems like the package was renamed, and that the new way to install & use it is
pip install retrolab
jupyter retro
@jtp can you please confirm or inform this ?
Yes, and there is an announcement topic on Discourse:
@jtp what do you think about publishing another version jupyterlab-classic to PyPI/npm/both which would add a note about the rename in the README and maybe a console.log/logging.warn message to inform existing users of jupyterlab-classic · PyPI?
Sounds good.
Would you mind opening an issue on GitHub - jupyterlab/retrolab: JupyterLab distribution with a retro look and feel 🌅 to track this? Thanks!