Re: Official plugin manager?

[dana <dana@dana.is>]

On 2 Jan 2020, at 05:37, Sebastian Gniazdowski <sgniazdowski@gmail.com> wrote:
> Okay. I think that the one interesting point was the 3rd one. I think
> that it harms Zshell's picture out there by making the following
> thinking common: that you cannot use Zsh without loading a bloated
> framework like Oh My Zsh.

I agree that zsh provides an underwhelming out-of-box experience for novice
users and that they are funnelled into OMZ in an attempt to address that. It'd
be nice if that wasn't the case. I'm not necessarily opposed to an official
plug-in manager, but i'm not sure it follows that creating one would solve
this specific problem.

Lots of zsh plug-in managers already exist. Any one of them can be used by
blog authors to 'share decent zshrcs which would also include the other needed
setopts and zstyles, etc.' *right now*. The existence of OMZ does not prevent
this. OMZ itself can be used to manage settings like that. The only thing that
would be different about a first-party plug-in manager is that you wouldn't
have to install it first. Beyond that, for *managing plug-ins*, it makes very
little difference.

Novice users don't use OMZ because they want to manage plug-ins. They use it
because they want a command they can just run to get the fancy completion and
prompts they see in screen-shots on dev.to or whatever, and OMZ fulfills that
desire by unconditionally modifying the shell's configuration to enable
whatever its developers consider desirable functionality. I assume that's what
you meant when you said that OMZ is 'bloated'. It *is* bloated, but that's
exactly the selling point. Were we to provide a first-party plug-in manager,
in the absence of other changes, it would simply be used to install OMZ, or
something like it.

When i first started using zsh, i had the same experience with the new-user
wizard that Roman did, and it's my impression that that's not uncommon. It's a
good idea in theory, but I think the success of projects like OMZ and fish
shows that most users now don't really want the degree of control that it
provides. They just want the cool stuff. If that's true, i feel like having
the wizard simply offer to turn on that (pre-determined) cool stuff would be
an easier and more effective way to cut into the OMZ use case than a plug-in
manager.

dana