Hello, following the discussion in the latest community meeting, I am opening an issue to solicit feedback on what would be the best way to name notebook-related resources in the UI. It would be ideal if we could follow a uniform approach and have a unified experience throughout the Kubeflow UI, so I think this discussion is of interest to @avdaredevil and @jlewi .

Here are some starting thoughts, based on the discussion in the community meeting:

• The New Notebook UI [issue Jupyter notebook manager UI that uses the new CRD #1995 , PR Jupyter UI that manages Notebook CRs #2357, by @kimwnasptd and @ioandr ] uses @lluunn 's Go controller to create pods that run Jupyter inside them.
• We don't want to be referring to "pods", and "namespaces" and other K8s entities in the UI. It's best to talk in Data Scientist-friendly terms.
• There's two different things that we call "Notebooks":
  • The actual notebook file (*.ipynb), that the data scientist opens
  • The actual server/environment that the user needs to spawn and to which they connect, so they can open their notebooks. We run this as a pod, on K8s.
• The new UI allows multiple notebook servers / pods to run at once.
• Running multiple servers is important so users can have multiple things running in parallel, perhaps in different namespaces, so they can collaborate in different teams.

So, here is a step-by-step description, with two alternatives:

1. The Notebook UI is accesible as Notebooks from the central dashboard. If I am a data scientist, and I want to find myself inside a notebook to start working, and I see a link called "Notebooks", I will follow it.
2. I am shown a list of my notebook pods/servers. I have no notebook servers, so I click on ➕ to create my new notebook server. I can have multiple notebook servers. My notebook servers have a name, so I can actually tell them apart. [Note there is no reference to "K8s" or "pods" at this point]. My notebooks run inside my notebook servers. [There's at least two different options of what to call items in this list, more below].
3. Once my server is done [I can see it changing going various transient states until it becomes a nice green Running], I can connect to it.
4. Once connected, I find myself inside a familiar environment, and I can open my notebooks / *.ipynb files.

So, there's two alternatives for step (2), none of which can be 100% precise, so we have to compromise:

• Option A: The list says "Notebooks ➕ " at the top. The label at the top of the list is simpler, "Notebooks", but there is the risk of confusing the actual notebook files with the servers to which I connect to open said files. Naming/renaming things in this list of servers does not rename my actual notebooks files. And if I connect to a "Notebook" from the list, why am I then shown multiple notebooks files? I think this is the argument raised during the community meeting, please chime in.

• Option B: The list says "Notebook Servers ➕ " at the top. The list is an explicit list of servers, not notebook files, the names refer to distinct servers, and I can connect to them, to see the notebook files that live inside them. But then, the drawback is that the left-hand side item I clicked on was "Notebooks", and I was actually shown a list of "Notebook Servers" when I selected it.

So, what do you think would cause the minimal confusion?

If we allow the slight discrepancy between what's shown on the navigation panel ("Notebooks") and what's shown as a title to the list ("Notebook Servers") I prefer option B: it combines short, simple words in the left-hand side navigation panel that draw attention to the different functional parts of the platform (Pipelines, Notebooks, etc.), while still being explicit with what is actually shown inside the list ("Notebook Servers", not "Notebooks" themselves).

Looking forward to any feedback you may have!