Re: [Jmap] I-D Action: draft-ietf-jmap-sieve-02.txt

[Ned Freed <[e-mail]>]

I think there's a more fundamental question we need to ask here: Is the
managesieve model the right one for handling sieves in JMAP?

By "managesieve model" I'm referring to the underlying data model where
there are multiple named sieves and one active sieve.

My specific concern is interoperability between clients that fully support this
model and clients that don't. A client that doesn't want to offer the ability
to have multiple sieves is likely going to just present the active script and
let you mess with it. Things can get interesting if another client that does
support multiple scripts comes along and switches to another script.

Of course there are ways of handling this, some requiring protocol/server
changes and others requiring extra vigilance on the part of clients. But before
we start exploring those - and given the ratio of single script to multiscript
clients I've seen I don't really think we have a choice - I think we need to
ask if multiscript facilities are something we want.

				Ned

P.S. I speak as someone who has implemented managesieve server side without any
real difficulty, and forsee no difficulty doing the same in JMAP. This is all
about client needs.