If it meant I could keep the front-end (shell) semantics but work with
slightly different back-end objects, I would at least be able to exploit
muscle memory.

in all other regards I think you're right. Fuse FS type overlay would make
this entirely moot.

I remember a few decades ago having similar feelings about "can we map MH
onto IMAP" and the retorts were "yes, but you do it first" along with "no,
and it's a stupid idea" and I think this "can we make MH work with more
modern <x>" is the hisenberg problem of our age: we love the s/w but the
rest of the world keeps on inventing things. STOP INVENTING THINGS!

cheers

-G

On Mon, Feb 16, 2026 at 1:33 PM Ken Hornstein <[email protected]> wrote:

> >What do people think? Could this work for a more natural (n)mh workflow
> >with mail?
> >
> >https://datatracker.ietf.org/doc/html/draft-ietf-jmap-filenode-04
>
> I read this, and it seems like it's describing "NFS using the JMAP
> protocol".  Which ... does not seem to change anything from the nmh
> perspective?
>
> If I'm wrong about this protocol, please, explain it to me!
>
> --Ken
>

Reply via email to