I asked on Slack and the immediate responses were:

• Use lsst-scipipe-abcde because that's what newinstall has used for a year and people are already accustomed to it.
• Use scipipe-abcde because lsst doesn't really matter and it's not like you are going to clash with other conda envs any how.
• Use rubin-abcde and declare that science pipelines is driving the conda environment for the rest of the observatory (which in a way is exactly what is happening).

And of course newinstall would be changed to match whatever. I'm edging towards the third option.

Talked to Leanne about it; the quickest, safest, easiest thing to do is lsst-scipipe-8hexchar. We can do a global search and replace for the prefix string, but I'd prefer to save that for when we get all the third-parties into conda. So I think this ticket, already reviewed, should be merged as-is.

What will remain is to have the lsst-dev shared stack match this result; I'm not exactly sure why it doesn't right now if newinstall already has the hash appended.

I'm not exactly sure why it doesn't right now if newinstall already has the hash appended

That one's easy: the shared stack setup script explicitly overrides it.

Why? Because it wants to be able to execute commands in the stack environment without having to source other scripts to set things up. It's not pretty, but it works.

Re-reviewed the further modifications.  Generally fine, although there appear to be excess PATH entries for miniconda that should be removed, and I think csh is broken.  The fix is non-trivial, however, and could be postponed to another ticket (that might be a way to judge if csh support is still required...).

I think given the comment from yesterday you can put this ticket straight into Reviewed.