Proposal to start a new implementation of Thunderbird based on web technologies

Paul D. Fernhout pdfernhout at kurtz-fernhout.com
Tue Apr 4 12:44:31 UTC 2017


Thanks so much for pointing out JMAP! JMAP looks like a great option for
moving forward towards any web-technology-based UI for Thunderbird. JMAP
seems especially well-suited for mobile devices which hopefully implies
good support for the desktop need of scrolling through thousands of
messages in a folder.

The biggest short-term benefit from using JMAP may be if we consider
flipping around the client/server idea of Thunderbird "supporting" JMAP.
Instead of Thunderbird first being a JMAP client to a remote mailstore,
Thunderbird could first be a JMAP server to a web-based UI client.

With that initial approach, the UI connection to the Thunderbird core
would benefit from all the great JMAP design as an email-domain-focused
communications protocol. Otherwise, we would be essentially be
reinventing JMAP when we think about how a web-based Thunderbird UI
would talk to a Thunderbird server.

I have rough ideas on a more general API based on other work (like I
experimented with in Pointrel/Twirlip versions) -- but JMAP is a much
lower-risk strategy to start with and could fit well with the
email-centric Thunderbird culture.

And, if we need more features (say, to make Thunderbird into a more
general federated knowledge-sharing platform -- which is my ultimate
hope for Thunderbird's future), we could perhaps suggest improvements to
JMAP.

And JMAP as-is might be good enough anyway to build other things on top
of -- in the same way IMAP has been used for other more general purposes
in Kolab. See:
https://en.wikipedia.org/wiki/Kolab
"The special idea behind Kolab is the use of IMAP as an underlying
protocol not only for email, but also for contact and calendar entries.
These entries are simply saved in special IMAP-folders utilising the
Kolab XML-format and the IMAP-server takes care of the storage and
access rights."

Here is another post I made has more details on why adding a JSON/JMAP
server to Thunderbird could make sense to manage risk by supporting an
incremental migration to web technologies:
https://mail.mozilla.org/pipermail/tb-planning/2017-April/005357.html

--Paul Fernhout (pdfernhout.net)
"The biggest challenge of the 21st century is the irony of technologies
of abundance in the hands of those still thinking in terms of scarcity."

On 2017-04-03 12:09 PM, Tanstaafl wrote:
> On Mon Apr 03 2017 12:03:22 GMT-0400 (Eastern Standard Time), Gervase
> Markham <gerv at mozilla.org> wrote:
>> Why is POP and not IMAP in the MVP? I'd say if you are going to do one,
>> IMAP is the modern mail access protocol.
> 
> Actually, I suggest that IMAP is no longer the 'modern' one. While I am
> not seriously suggesting the new TBird shouldn't support IMAP, I suggest
> JMAP support be made a mandatory criteria. JMAP is designed to provide
> all of the advantages of IMAP, while addressing most if not all of its
> shortcomings:
> 
> http://jmap.io/
> 
> Both Cyrus (3+) and Dovecot (2.3+) are primed to support it, and there
> is a Thunderbird bug to add support for it:
> 
> https://bugzilla.mozilla.org/show_bug.cgi?id=1322991



More information about the tb-planning mailing list