jupyterhub / nbgitpuller

Jupyter server extension to sync a git repository one-way to a local path
https://nbgitpuller.readthedocs.io
BSD 3-Clause "New" or "Revised" License
210 stars 86 forks source link

Rename `nbgitpuller` to something more descriptive of new functionality #233

Closed choldgraf closed 2 years ago

choldgraf commented 2 years ago

Proposed change

Now that we are introducing a lot of nice new flexibility for this tool (see #194 ), we might consider renaming it to be more descriptive. The name nbgitpuller will make people assume that we only pull git artifacts, whereas we can now do more than this!

Benefit

The main reason for doing this is to make it easier for potential users and developers to understand the scope of this project. By removing the explicit call-out to git, people might more quickly imagine using this tool for more workflows like Google Drive and Dropbox pulling.

(Optional): Suggest a solution

A few options that were thrown around in a recent meeting (some more serious than others)

ryanlovett commented 2 years ago

I suggest jupyter over nb, since it isn't notebook specific. I think nb comes from the days when most notebook and notebook server extensions started with that.

schwag, because it is usually a bag of stuff passed to you just when you're getting started at some event. Not really serious though. :)

choldgraf commented 2 years ago

I vote for Doug

(good point about not being notebook specific, I believe @manics brought up the same point)

If not Doug, then:

what about jupyter-content-puller? I like jupyter-data-puller but worry it will give people misconceptions about the workflow that this tool was designed for (which is distributing content to hub users). I feel like we generally recommend other workflows for pulling in data into Jupyter than this

ryanlovett commented 2 years ago

jupyter-content-puller is a good one. I think of nbgitpuller primarily by what it does (pulling) rather than its purpose (pushing, distributing) but maybe that's not how it should be named.

jupyter-payload-puller invokes jupyter's spacey theme.

Suggesting names is fun.

yuvipanda commented 2 years ago

I think nbgitpuller is like... the 4th or 5th name for this project? It's gone through a lot of renames, but nbgitpuller is the one that has stuck the most. More importantly is the idea of an 'nbgitpuller link', and changing names here would require re-education of that too - people would have to remember that 'oh, it is a jupyter-content-puller link, but that is the same as this nbgitpuller thing'. data8 still calls them 'interact links' (interact is a former name), and that's still confusing. So I want to say that renaming has fairly high costs.

/cc @ericvd-ucb, who has experience explaining nbgitpuller links to lots of users. Would a rename be confusing?

Internally, we still use git and there's a core feature we provide that relies on that - the merging of pulled content with what's locally there.

My intuition is to avoid yet another rename...

yuvipanda commented 2 years ago

Names it has had so far (I'm looking at history of setup.py for this):

  1. nbpuller
  2. gitautosync
  3. nbautosync
  4. interact (although I think we threw away that codebase?)
  5. ???

6b56b46a381f157d5f720fbec07bb070e5e7507f was when the nbgitpuller name was adopted.

choldgraf commented 2 years ago

I agree that renaming is a disruptive change. I guess the question is whether we think the current name is misleading enough to new users that it justifies changing and disrupting the workflows of existing users. It's hard for me to know the answer there without doing some user research :-/ .

There were a lot of nodding heads when we discussed this in the team meeting, which is why I opened this. But if people in general don't feel very strongly that we should rename, I don't think we should push it.

Does anybody want to vote in favor of "we should change the name now"?

consideRatio commented 2 years ago

Note that even though we pull from ZIP files etc, we still use git in the update process. If you pull a zip file twice, files are updated like if it was a new HEAD in a git branch -- I think.

So, if we remove git entirely, we should be clear that it has a big role no matter what.

yuvipanda commented 2 years ago

I'd love to hear from @ericvd-ucb, as he's a lot of experience with evangelizing this.

ericvd-ucb commented 2 years ago

Hi - Im not sure I have much novel to contribute -

choldgraf commented 2 years ago

Thinking back on this, it feels like we are in this situation:

  1. nbgitpuller isn't the greatest name, but it's not that bad either
  2. Many communities have already used the name nbgitpuller for several months now, and changing it would be disruptive

To that extent, I feel like we should just leave the name as-is, and focus our efforts on improving the use-case documentation via things like https://github.com/jupyterhub/nbgitpuller/issues/236 . Do others agree or object?

consideRatio commented 2 years ago

I agree on leaving it as it is currently, because I feel no name is a great fit anyhow, and the extended functionality we may adjust to is optional rather than built in.

yuvipanda commented 2 years ago

I'm going to close this as 'let us not change it'. Thanks for the discussion everyone!