LibreTexts users are confused about the "Restart" button in the binder plugin

[moorepants]

See this explanation of the confusion in this video: https://youtu.be/gA9s0NQRVzY?t=1943

I thought the restart button, restarts the kernel for the page. That way if you lock up the kernel, you can restart it. But maybe it isn't working as expected.

[sandertyu]

Playing around with the restart button, we've found that it basically acts as a "reset" button for the page. Because data generated in one code cell will persist across the page, the restart button will erase all that data by starting a new kernel as if you had refreshed the page. However, it's much simpler and cleaner than simply restarting the page, as it doesn't wipe any of your code changes and it neatly shuts down the current kernel.

For instance, if you have a code structure where cell 1 has a = 10 and cell 2 has a + 5, cell 2 will give an error if you run it without running cell 1 first because a has not been defined within the runtime of the current kernel. If you click the restart button of either code cell, then you must again run cell 1 in order to not receive an error when running cell 2. So it simply restarts the kernel which exists sort of implicitly across the entire page.

Now that we know this, what exactly would you like us to do. Is there documentation we should add to?

[moorepants]

Everything we want to communicate to LibreTexts users about the ckeditor plugin and the jupyterhub should be added to this tutorial:

https://chem.libretexts.org/Courses/Remixer_University/LibreTexts_Construction_Guide/05%3A_Interactive_Elements/5.02%3A_Jupyter_Notebooks_(Executable_Programming_Code_and_Figures)

either as additions to this page or as new pages in the construction guide.

[sandertyu]

Okay, I'm thinking we should devote an entire page to explaining how exactly the code integration process works. I only fully realized this recently, but it makes a LOT more sense to understand how the code cells work if you think about it as running a jupyter notebook in the background (and this is basically what is literally happening as far as I know). Each code cell represents a notebook entry, and they can be run independently of eachother. Data will persist across the entire webpage unless the kernel is restarted, exactly as it would in a notebook.

I think we could repurpose that page you linked there to explain this functionality, and then basically copy paste that page's current contents onto a new page specifically for using ckeditor plugin.