In that case, I’d recommend some screenshots to be sure everyone is on the “same page”.
Here’s what I see when I open the Anaconda Navigator.
Here’s what I see when I click on the “Launch” button for Jupyter Notebook from Navigator. Jupyter starts in your “User” folder, which is likely the cause of your comments about Jupyter’s file interface, which forces one to click individual folder icons to get to the desired location.
An alternative method to launching Jupyter is with the “Anaconda Prompt”, which on Windows is just the cmd
prompt with conda settings initialized. Using this method, one can see what conda environments are available with conda env list
, then activate one besides (base)
with conda activate <env-name>
, then one can “pre-navigate” to whatever folder is desired and launch Jupyter with the command jupyter notebook
. Below is what it looks like on my Win10 machine.
And here is the resulting Jupyter notebook session, which opens in the empty “Example_Folder.” Notice that there’s no difference in functionality or capability of Jupyter due to the different launch mechanism.
Since Anaconda Prompt is a part of the “Anaconda Ecosystem,” I don’t think of it as separate from Anaconda, but it is also not launching Jupyter from the Anaconda Navigator GUI.
Something that *is* separate from Anaconda is the Window PowerShell command line. But Jupyter and conda can both be utilized from there, as well. Below, conda is used to change the virtual environment, then the directory is changed and Jupyter is launched.
And here is the resulting (nearly identical) Jupyter session.
For what it’s worth, I almost never use the Navigator. In fact, these days I use VS Code for most of my work, including the Integrated Terminal pane, which can be PowerShell or any other shell that is installed. For example, the Windows Subsystem for Linux bash shell is available and can perform the exact same functions as Anaconda Prompt and PowerShell. See below.
And the resulting Jupyter session (the only thing changing is the port number as multiple Jupyter sessions are created.
The main “takeaway” point I’m trying to get across is that the use of conda environments is not tied in anyway to the Anaconda Navigator GUI, and in fact the Jupyter notebook sessions are all running a version of Jupyter installed as part of the Anaconda distribution. If you look at all the shell sessions above, Jupyter is loading from ~/Anaconda3/envs/py36/…
So, making the statement “Jupyter Standalone Might Just Be Better Than Anaconda” could confuse folks new to Jupyter and Anaconda, which I believe is your intended audience.
I believe a more accurate statement would be “Launching Jupyter From The Command Line Is More Powerful Than Relying On The Anaconda Navigator GUI.”
I applaud your impulse to share your findings, and I encourage you to continue to do so. Keeping terminology unambiguous is a universal challenge that we’re all faced with, though!