Jérôme Dockès

> (in this PR) rename `scale_bg_map` to `bg_scale_map` I don't love `bg_scale_map` -- I think it's harder to guess what it does than `scale_bg_data`. but I do see the value...

> Besides, I think allowing to set `alpha` values for surface maps could help pave the way towards being able to plot several surface maps at the same time (cf...

> I agree that setting transparency for background maps probably isn't useful and that we could deprecate this parameter. > Moreover, maybe we could remove `darkness` (ie `bg_darkness`) altogether and...

>> I think so, as long as we make sure plots look good with possible default choices such as fsaverage sulcal depth or curvature sign. > > Yes I agree,...

As mentioned by @bthirion in both linked issues, for the case of `plot_stat_map`, using `plot_img` instead does suffice. ```python from nilearn import plotting, datasets plotting.plot_img(datasets.load_mni152_template(), colorbar=True, vmin=.5) ```  For...

The screenshot uses `view_img`. `plot_img` always allows setting any `vmin`: ```python from nilearn import plotting, datasets img = datasets.fetch_neurovault_motor_task()["images"][0] plotting.plot_img(img, colorbar=True, vmin=-.5) ``` 

we could remove the check that forces the cmap to be symmetric in view_img when there are negative values, as symmetric_cmap=True is the default

bthirion ***@***.***> writes: > I think that this comes from historical reasons, but I indeed think that `plot_img_on_surf` shoudl accept a vmin argument. I agree. ATM it relies on `plot_surf_stat_map`,...

> I can imagine this subject is often being discussed... > I personally enjoyed contributing to repos which implemented some `pre-commit` logic with `black` for instance (I know it has...

but regardless of whether the tools have improved, the issues pointed out in this issue and in #3317 and #3308 (one more thing to ask of contributors, one more thing...