Default e-mail client

Registered by Jason Warner

Consider switching to Thunderbird as default email client.

Blueprint information

Martin Pitt
Chris Coulson
Chris Coulson
Series goal:
Accepted for oneiric
Milestone target:
milestone icon oneiric-alpha-3
Started by
Martin Pitt
Completed by
Martin Pitt

Related branches



Work items for oneiric-alpha-1:
[chrisccoulson] Review space needed by Thunderbird: DONE
[pitti] Final answer on CD space constraints, Thunderbird on CD or DVD: DONE
[andreasn] Meet with john lea to get general design direction: DONE
[johnlea] Provide general guidance to andreasn on design direction: DONE
[chrisccoulson] Implement indicator-appmenu integration: DONE

Work items for oneiric-alpha-2:
[chrisccoulson] Evaluate possibility to share libxul between FF and TB on the CD (even though the 6-week releases won't be possible): DROPPED
[andreasn] Convert initial list of desired styling improvements into bug reports and mockups: DONE
[chrisccoulson] Establish contact with Lightning developers for calendar integration: DONE
[chrisccoulson] Provide a list of necessary tasks for launcher and messaging menu integration: DONE
[chrisccoulson] Evaluate current Lightning integration and identify needed improvements: DONE
[chrisccoulson] Evaluate upgrade reliability and release strategy for enigmail, Lightning, and other extensions: DONE
[mconley] Initial work on contacts integration, demonstrate read capability on contacts: DONE
[jasoncwarner] Initial evaluation of work progress: DONE
[chrisccoulson] change seeds to add tbird and drop evolution: DONE

Work items for oneiric-alpha-3:
[seb128] Discuss strategy for creating new Ubuntu One address book (on new install): DONE
[allison] Discuss strategy for creating new Ubuntu One address book (on new install): DONE
[chrisccoulson] Discuss strategy for creating new Ubuntu One address book (on new install): DONE
[charlinepoirier] Usability testing, compare Thunderbird and Evolution: DONE
[mconley] Deliver launcher integration: DONE
[chrisccoulson] Deliver messaging menu integration: DONE

Work items for ubuntu-11.10-beta-1:
[chrisccoulson] Deliver calendar events in the datetime menu: DROPPED
[chrisccoulson] Merge indicator-appmenu integration upstream: DROPPED
[chrisccoulson] Evaluate possibility to save space in TB packages on the CD: POSTPONED
[mconley] Final work on contacts integration, ability to read/add/edit contacts in existing address book: DONE
[jasoncwarner] Final evaluation, will Thunderbird be default email client: DONE

Work items for ubuntu-11.10-beta-2:
[andreasn] Review styling improvements with John Lea: DONE
[johnlea] Review progress in styling improvements, provide guidance on further improvements needed: DONE

Work items for ubuntu-11.10:
[chrisccoulson] Review update plan for extensions with security team: DONE

Rejected for Oneiric cycle:
- Overlay scrollbars are "nice to have" for 11.10, will be required for 12.04. (This work will be done for Firefox anyway for 12.04, and Thunderbird will benefit from that.)
- Won't change Thunderbird packaging to use Firefox XUL (FF and TB release cycles are 1 week out-of-sync, which is a substantial portion of their 6-week release cycles).
- For 11.10, we won't expect Thunderbird to create new Ubuntu One address books. Need to plan how Ubuntu One will create them.

chrisccoulson, 2011-05-25: I've deleted all the bug links, as I'm not sure who added those. Linking bugs to blueprints makes them appear as tasks assigned to me on the work item tracker, and shouldn't be used as a way to bring particular bugs you want to see fixed to our attention

AJenbo, 2011-05-31: Thunderbird not automatically installing it's locals will also have to be adressed.

pitti, 2011-05-23: Can you please clean up the description and whiteboard for drafting, and only keep the actual WIs? Please note that all linked bugs will automatically be regarded as work items, so we only need the additional ones in the whiteboard.

pitti, 2011-05-26: CD size constraints: will allow us to free a couple of MB less than we'll need to add with unity-2d, Qt, and LLVM in the current graphics drivers; we can get back the rest with some better squashfs construction in live-build. So essentially every new feature needs to have a balanced budget now; i. e. for everything you add you should also propose something else to drop instead. Assuming that we will keep e-d-s, evolution uses about 6 MB (compressed package size), while current Thunderbird uses 12 MB; Chris said that the daily builds use about 20 MB; i. e. we would need to free 14 MB by cutting other things or making tbird itself smaller (the 8 MB growth can't be entirely justified?). Note that this delta does not yet include a calendar replacement, which Evo provides, but Thunderbird doesn't.
Also, can we re-evaluate sharing libxul between firefox and tbird at least for the CD builds? Shipping two huge libraries which are by and large identical seems like a huge waste?
Wrt. the DVD, we still shouldn't waste space there (cf. big library duplication), but it hurts a lot less there, and adding 10 MB or so is not a problem.
I'd appreciate some WIs to investigate sizing down the tbird package, otherwise this looks fine to me as a plan.
Sharing for the release and reintruce it in -updates (10 MB), and removing half of the fat that the package grew since natty final (8MB), we'd have the 14 MB back, and we would be good. Is that realistic?

Further comments from this whiteboard are saved in the Etherpad doc:

Evolution pro's:
- Good integration with the desktop already (eg, messaging menu and appmenu)
- Integrated with existing translation infrastructure
- Calendaring functionality by default, and integrated with the desktop
- Support for syncing contacts with U1
- Contacts sync with GMail
- GNOME release process is better aligned with our 6 month cycle
- Exchange support (no idea how well this works, but it exists)

Evolution con's:
- Outdated and confusing UI
- Historically has been fairly slow and unstable (although it is better now)
- UI is pretty bad on netbooks and other small form-factor devices
- I'm not convinced that Evolutions additional features are that important to our target users

Thunderbird pro's:
- Responsive and more active upstream
- Familiar brand for users moving from other operating systems, which has the same benefits as shipping Firefox
- Lots and lots of extensions, and a very rich extension framework
- Initial account setup is so much more intuitive
- I like the tabbed interface ;)

Thunderbird con's:
- Translations not integrated with Launchpad (we have the same issue with Firefox though)
- Integration with the desktop lacking (no messaging menu or appmenu)
- No exchange support (not true for lightning: there is a working lightning data provider, )
- Calendaring support is only available via an addon (Lightning), and is not integrated with the panel clock
- No GMail or U1 contact sync (although GMail contact support is available via an addon)

That is a brief summary of where we were at the last UDS. Since then,
the following work has happened:

- I have written an extension to add appmenu support to Thunderbird (and Firefox), which will be merged upstream at some point.
- Mike Conley has been working on integrating Thunderbird with the messaging menu and the launcher in Unity. There's pretty good progress on this already, with extensions already available to test.
- Mike has also started working on U1 contact sync.

However, there have also been changes to the release schedule for Firefox. It's currently unclear how these affect Thunderbird.

So, as of today, these are the things that I think we would miss if we switched the default client to Thunderbird:

- GMail contact support - I think this one is fairly important, and I'll offer to help get this fixed if I can.
- Calendaring support - Whilst I agree that the existing calendar integration in Evolution is nice, I'm not convinced that it's a deal-breaker for our target audience. I use the calendaring in Evolution for work, but I don't miss it at all on my home desktop (where I use Thunderbird).
- Exchange support - again, whilst Exchange support may be important in a corporate environment, I don't think this is particularly important to most users we are targetting. This seemed to be the consensus amongst the people who attended the session at the last UDS too.
- Translation support in Launchpad - I'm already planning to work on this anyway.


Work Items