Re: [Nmh-workers] nmh architecture discussion: format engine character set

[Anthony J. Bentley]

Hi,

Earl Hood writes:
> Some general comments about this discussion:
> 
> - Message should be stored in their original forms.  I.e.  The
>    character encoding transformation should only be done for
>    display/access purposes.
> 
>    Main reasons:
>      - Protects against bugs in encoding transformation code.  The
>        original message is always left untouched.
>      - Maintains compatibility with folks that use external tools and
>        scripts to access nmh messages.
>      - Avoids invalidating any digital signatures since any
>        modification of the message will invalidate them.
> 
>    If there is a desire to support transformation of incoming mail, so
>    mail is stored in "normalized form", then it should be a configurable
>    option.  I personally would not want all my stored email converted to
>    UTF-8 due to reasons cited above.

Not to take things too far off-topic, but in my wild imaginings I've
wondered how practical it would be to implement a mail client that
used a mh mail store as a base to expose a filesystem (maybe over FUSE
or 9P) that contains individual messages as decoded UTF-8 text files,
and sub-hierarchies as necessary for attachments/MIME, however that
would work.

It seems like a lot of room for error in implementation. But talk about
satisfying the Unix philosophy...

-- 
Anthony J. Bentley