Bryan Van de Ven

@bkos please provide a [minimal reproducible example](https://stackoverflow.com/help/minimal-reproducible-example) that can be run as a test case to verify. I'd encourage forcing one of the plots to have very large range to...

There are several possible ways this might be handled, so I have tagged as `discussion` for now. Adding new properties means adding the new behavior on top of all the...

Alternatively, we keep only a single numerical `axis_label_standoff` but add a separate "mode" property to dictate how it is interpreted. That's probably better than than the negative values approach. What...

> I actually want to merge this for 3.0. The status of CrosshairTool.spans will be experimental thought until a better solution emerges. Why not just wait, then? Providing users an...

> There are several reasons: Ok, well I am really only ok with experimental APIs if we also issue an actual warning of as some kind (similar to other projects,...

@mattpap I am inclined to revert this in the morning. I do not agree to having experimental features without loud warnings and we clearly need to decide an document a...

I have drafted BEP 10 to address policies around experimental features: https://github.com/bokeh/bokeh/wiki/BEP-10:-Policy-for-Experimental-Features

You're welcome to implement the changes that I requested, instead of ignoring a clear objection to merging in the first place.

> We have had experimental features for years.. The fact that we have acted too loosely in the past is not a justification to continue to do so forever. In...

> btw. .gitignore is currently badly set up (after PR https://github.com/bokeh/bokeh/pull/12368), which means you have to use git add -f path_to.png. Can you please refrain from asserting moral judgements on...