Vision for Thunderbird

R Kent James kent at caspia.com
Wed Jan 31 20:20:29 UTC 2018


Ben and I had a very long discussion about this issue today, 
particularly as it would affect the next technical direction for 
Thunderbird.

The question that I ask myself is, what would Thunderbird need to look 
like for it to be a major player in the email world, say with the same 
relative market share to Outlook and Gmail that Firefox currently has 
relative to Chrome and other mainline browsers?

This is particularly critical in the context of first technical steps 
toward a revamping of Thunderbird, which we hope to start very soon. 
We've generally agreed that a good context for that would be a rewrite 
of the address book, as it is both badly in need of attention, as well 
as sufficiently independent from the rest of Thunderbird that it would 
be a good experimental platform for alternate technologies.

What do people expect of their email client in the modern world?

*Big question #1: Overall scope.*

** I would say that the product needs to work reliably and consistently 
in two dimensions:

1)    Platform. Users expect to be able to access their personal 
information on a desktop platform (at least Windows and OSX, but we 
would undoubtedly also want to include Linux), on mobile devices (at 
least Android, iOS if possible), and from a web browser. (Currently 
Thunderbird users do this with different apps on each platform, but 
there are little inconsistencies in the behavior that makes that awkward 
and unsatisfying).

2)    Protocol. Users have a wide variety of primary sources for 
information, and they expect their email client to work well with all. 
That includes both open protocols (IMAP, POP3, and SMTP for email, 
iCalendar and calDAV for calendar, cardDAV for address book) as well as 
proprietary protocols, including the Google suite as well as Microsoft 
and Exchange Web Services. (Currently in Thunderbird, some protocols are 
supported in core, others are supported inconsistently and frequently 
unreliably in addons).

Is the goal for a future Thunderbird a product that satisfies both of 
these dimensions as primary (core) goals? Personally, I don't see that 
you can shoot for anything less if you hope to be competitive. (Whether 
we can marshal the resources to pull that off is another question, but 
related - if there is no vision there is no hope for the resources either).

*Big question #2: Security.*

Of course anything that we do would need to be immune to malware 
attacks, but the issue here is advances in communication pointed toward 
end-to-end encryption. There have been a variety of attempts to provide 
working products and protocols to this end, but this has not been a 
focus of Thunderbird. Should we re-prioritize our efforts toward 
accomplishing the goal of reliable, convenient end-to-end security? 
(There have been at least two third-party attempts to implement this in 
Thunderbird, which we so far have only hesitatingly encouraged. This 
priority could be accomplished in partnership with more commercial 
entity, but keeping in mind the backlash against such partnerships that 
Firefox has faced).

In addition to the end-to-end encryption issues, closely related issues 
hinted at by Ben are the issues of privacy and dependency resulting from 
dependence on central servers. Local storage on Thunderbird is not 
really prioritized currently as an alternative to server storage, and 
there are many details that would need to be attended to before that was 
really viable and reliable.

*Big question #3: Implementation polish.*

This may seem like an odd big picture question, but really a focus here 
would be in direct conflict, in the short term, with the other 
big-picture priorities.

There are many, many details of how Thunderbird currently works that 
really could use improvement. I'm sure everyone has their pet peeves 
that they would like to see improved. So the question is, should 
Thunderbird remain as primarily a desktop product focused on existing 
open protocols (IMAP, CalDAV, and CardDAV) and focus new efforts on 
performance, usability, and reliability within that restricted scope?

*Big question #4: Betting on Gecko.*

An issue that we have not come to consensus on is whether we believe 
that continuing life as a binary rebuild of Firefox is a viable future 
for Thunderbird. If we really wanted to follow Firefox, we would be 
moving into Rust and focusing on reliability and performance. Is that 
even viable? (I predicted two years ago that our upcoming release, 
Thunderbird 59 now 60, would be the last Gecko release we could 
successfully achieve.) Where someone stands on this technical question 
has largely dominated discussions of what steps to take next.

As we are currently contemplating a major investment in a future 
Thunderbird, I think that we as a community (or as potential Thunderbird 
Council members) need to have a clear understanding on where we stand on 
this. Without that, we have no clear direction and no vision that anyone 
could get behind.

*My approach*

Here's my viewpoint. In the short run, for the address book rewrite, we 
explore the technical issues around big question #1 Platform, so that we 
understand what are the challenges and tradeoffs involved in targeting 
all three platforms, as well as recommending a technical stack and 
architecture for future work. Based on the results of that, we make a 
concrete decision about future direction in these areas - but with the 
hope of being able to have a product that would support consistent 
workflows on desktop, mobile, and web platforms.  We end up with an 
address book that can work as integrated into existing Thunderbird, as a 
standalone desktop app, as a standalone web app, and as a standalone 
mobile address book. As part of that, (big question #1 Protocol)  we 
would support both open and proprietary protocols.

After that, we should take on a second major area for revision, with the 
likely candidate being the front-end thread pane view. Here the major 
focus is whether our choice of technical stack and architecture can 
deliver the performance and features we need for the future. This would 
also be integrated into the current product, with possibly a minimally 
viable prototype email client using that thread pane view.

At this point, my expectation is that a) Gecko would be increasingly 
difficult to maintain, b) the problems in retrofitting changes to the 
existing Thunderbird product would be understood as difficult and 
limiting, and c) we would have a clear alternate technology possibility 
in the address book and MVP email client. At this point we would decide 
to re-implement the rest of the critical functionality of Thunderbird 
into the future technical stack and architecture, eventually obsoleting 
the Gecko version of Thunderbird. This is a bet (big question #4 Gecko) 
that we could maintain Thunderbird on Gecko through a version 67 but not 
beyond.

Concerning big question #3 (Polish), the cost of this is that we would 
not be able to primarily focus on many small improvements to 
Thunderbird, but instead the visible improvements to the existing 
product would come from replacing two major components (the Address Book 
and email view) with more modern parts. Serious attention to polish 
would wait until it is implemented in the next generation of Thunderbird.

Concerning big question #2 (Security), these issues would not be 
prioritized, but reconsidered at the point of the full rewrite. PGP 
could reasonably be supported as a core capability, and if there are any 
other well established approaches as public protocols we could consider 
implementing. But this would not be an area we would be leaders in.

:rkent

On 1/31/2018 10:09 AM, Ben Bucksch wrote:
>
> Dear Thunderbird community,
>
> the Thunderbird project currently does not have a clearly defined, 
> written project manifest or vision in the form as the Mozilla 
> Manifesto <https://www.mozilla.org/en-US/about/manifesto/>. It would 
> be good to have something like it, but more concrete for Thunderbird. 
> We appreciate Thunderbird for specific qualities that it has, and I 
> think many of our reasons will overlap, but we never really defined it.
>
> We're kind-of at a cross-roads point with the project right now, and 
> it would be useful to have it written and clearly defined, so that we 
> can make sure the future strategy and developments maintain these core 
> qualities.
>
> So, I would like to start an informal survey about what about 
> Thunderbird is important for *you*, personally. Why are you here, why 
> do *you like* Thunderbird so much?
>
> Also, I'd like to know where you see Thunderbird in, let's say, 10 years.
>
> So, my two questions for you would be:
>
>  1. What specific qualities do you like about Thunderbird?
>  2. Where do you see Thunderbird in 10 years?
>
> It would be useful for you to give both keywords (to classify) and a 
> 1-line answer to know what you mean with that (because people mean 
> different things with "security", or there's a specific use case where 
> "performance" is important for you).
>
> Here's an example answer for the kind of answers would be helpful:
>
> 1. What specific qualities do you like about Thunderbird?
>
>   * Privacy is important for me. I use Thunderbird, because I don't
>     want to keep all data on central servers.
>     Keywords: Privacy, decentralization, local processing
>   * I want to keep my emails on my computer, so that I know I keep
>     access to it, even if I don't have Internet, or even if I change
>     applications, I want to know to be able to store my emails locally
>     in a standard format.
>     Keywords: local storage, standards
>   * I get 150 emails every day. Thunderbird filters help to keep on
>     top of them. It's also fast to open and read and move emails into
>     folders.
>     Keywords: Filters, UI, speed
>   * I have mailboxes with 30000 emails. Thunderbird allows to me
>     easily access them
>     Keywords: Speed, large mailboxes
>   * Security: I think Thunderbird does not have many security holes,
>     and I trust the project to quickly fix and properly any issues
>     that become known, not just paper over them.
>   * Encryption: I want to use PGP, and that makes no sense in webmail.
>
> 2. Where do you see Thunderbird in 10 years?
>
>   * I want Thunderbird on my Android phone
>   * I want to use WhatsApp from Thunderbird
>   * I want to share pictures with my friends easily and hassle-free,
>     and allow them to send me pictures and videos
>
>
> Once we have a good sample of responses from the community, we can try 
> to build a common picture. We should also reach out to a larger group 
> of end users.
>
> Ben
>
>
>
> _______________________________________________
> tb-planning mailing list
> tb-planning at mozilla.org
> https://mail.mozilla.org/listinfo/tb-planning

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/tb-planning/attachments/20180131/121d679b/attachment-0001.html>


More information about the tb-planning mailing list