> I assume you want to keep the email from each account separate?
Yes, that is correct.
> I've been an MH user going on 33yrs.
I'm curious: what did people commonly use for reading mail when MH was
just invented? Was it the Unix "mail" program?
mail for most folks. A few used emacs to do the job (mostly programmers given the vertical cliff learning curve for emacs).
By the time the early 90s rolled around (I think I switched to mh back around 93, perhaps 92), xmh, mailtool (in the SunOS world, anyway), pine, and mutt.
Those were what I was seeing in academia, anyway. [Mostly a Sun shop, but we had samples of most of the commercial UNIXes or UNIXish around at the time- DEC, NEXT, HP, AIX, whatever IBM had before AIX.]
> > Using different UNIX accounts ensures 100% separation. You can do
> > everything under one ID in theory, but you'll spend a lot of effort/time
> > switching email IDs via different profiles. My opinion is that this will be
> > error prone unless one has a lot of self discipline.
>
> And I'll second the suggestion. It is the easiest, cleanest solution and
> avoids any possible confusion where you sent a work email via a personal
> account *or* sent something personal via your work email.
I understand that this ensures that the accounts stay separate, but
managing multiple user accounts is not exactly light work. I guess the
use of separate UNIX accounts may be appropriate for particular use cases,
but I do not need such a strict separation at the moment.
Having managed multiple UNIX accounts in multiple domains (plus a shared root account), the effort is not that high for multiple accounts. If there is a shared computer or domain, you can pretty easily share ~/bin, dot files, and the like without much pain (did this for years). Throwing in X applications as both users being active at the same time is a bit harder if you wish to do so securely (no "xhost +" stuff). With a separate work and personal environment on different hardware and domains, yeah, that is high effort to keep up to spec. That is also the case where you probably _want_ a strong separation however.
Trying to keep my nmh stuff straight with 2+ mailboxes all ending up in one data store and with one set of nmh configuration files is hard and is going to set you up for regression issues (change that is needed for $work breaks stuff for $me). It is much harder to do that with different user IDs.
At the risk of scope creep, no matter which path you pick, give thought to using some form of version control (the archaic RCS or SCCS are up to this job, git or subversion would presumably be as well, there are many others to choose from as well, use simple copies somewhere else if nothing else). You will be tinkering for a while to get things to your liking. You will make errors and you will wish to roll back to that works less than optimally version in place of the doesn't work at all versions.
Disclaimer: programmer by formal training, sys-admin by vocation, so my use model and skill set are not the most common for nmh use.