On 27/01/05 12.54, John J. Foster wrote: [snip] > a decent addressbook How about this for an idea (maybe not 1.6.0 material, but it might be cool longterm): Cut away the addressbook! What I'm thinking is that the alias features could be refactored into a seperate executable (which could be fast because it does not have to do much), and the places where mutt interacts with it could be replaced by userdefined commands, like crypto support. Then a "decent" addressbook could be written, using whatever backend one found desirable (maybe one for the evolution-data-server interface, one for ldap, one for searching online addressbooks, etc). It might even be possible to write an "addressbook multiplexer", that would search a given list of addressbook-implementations (say mutt-style, evolution, ldap, web) in order and add new entries to one or more of them (possibly presenting an extended interface). As far as I can tell the number of features needed in such programs is rather small: Add a new entry (here it might be nice to support invokation of a (text-based obviously) gui, for additional data entry. Lookup an address, providing a list of possibles (mutt could still show the list and do selection, or that could be out-sourced). Maybe there could be an option to invoke a gui search (text-based again), to allow more advanced/plugin specific searching. Efficiently map from addresses (and perhaps a few other inputs) to names (etc.). This might be done by running the program as a pipe, with mutt holding both ends. Perhaps with a commandline argument fallback for prototyping/scripting. What else would be needed? I know this is a rather big solution, but it would make mutt simpler, while allowing for infinetely better addressbook features. Just another late night case of thinking out loud. /dossen
Attachment:
pgpI3jtnMxImh.pgp
Description: PGP signature