Thunderbird Thoughts

I’ve just noticed Mitchell’s and mscott’s posts on the future of Thunderbird. I’m glad this discussion is happening. Here are my thoughts.

I note several comments in Mitchell’s post saying “Well, hire more developers”. It seems to me from what she’s saying that money is not the problem – management overhead is. (Of course, “Thunderbird takes up too much management time worrying about what to do with it; what do we do?” is a self-fulfilling prophecy.) Anyway, throwing money at the problem won’t solve it. Having said that, if more money would help whatever the new plan is, I see no reason the Mozilla Project should not provide it.

Commenters have also brought up the Mozilla Manifesto. The key thing to note here is that we can’t do every activity which is compatible with or furthers the goals of the Manifesto. For example, it’s important for internet openness that the SIP open standard wins the VoIP protocol battle. But that doesn’t mean we should be writing a VoIP client. Having said that, we already do email, which counts for a lot.

So what about Mitchell’s options?

Option 1 makes no sense on several levels. Firstly, why start a Thunderbird Foundation when we have a perfectly good Foundation already looking after it? And even if a separate entity was needed, the Software Freedom Conservancy could provide all that was necessary without the effort of setting up our own new non-profit. (Maybe that’s Option 1b.) The SFC option counters a lot of mscott’s objection to Option 1.

Option 2 talks about keeping Thunderbird development within the Foundation, perhaps with a new Corporation-like subsidiary. I think the choice between this and option 3 (independent consulting) ultimately depends on how the two or three core developers feel they would like to operate.

Of this option, mscott says:

This option doesn’t satisfy the goals of either the Mozilla Corporation or the new company, since a Thunderbird subsidiary will not further the Foundation’s mission to advance the open web, and will continue to draw resources away from Firefox.

But I would respond: since when did the Foundation have a mission to “advance the open web”? Last I read, the mission was to “promote choice and innovation on the Internet”, and the Internet is explicitly not the web, but a superset of it which includes email. Also, whether this satisfies the goals of the Mozilla Corporation is an irrelevant concern, given that their personnel and management would not be involved with the new venture. For the same reason, I don’t see why this option would draw resources away from Firefox. I think mscott has short-changed Option 2 somewhat.

I understand the desire to have an organisation fully focussed on Firefox. The Mozilla Corporation could be and is that organisation (with our without a name change). But the Mozilla Foundation should not and need not be. We are here to take the wider view, and can look after the projects in our care without taking up MoCo management time. If there are people who want to be focussed on Firefox but are being distracted from that by MoFo duties, then perhaps we should consider making changes to their roles, rather than make MoFo another Firefox-only organisation.

Option 3 is the consulting company option. This has the advantage that the market will quickly tell the core Thunderbird developers what it wants – answering, for example, the “pure email” vs. “Outlook competitor with calendaring” question. But there’s a higher risk of failure, as if no-one wants to stump up money for Thunderbird, then the core developers will need to start working on something else to pay the rent.

I think the right move here is to set both the Corporation and the Foundation free to pursue their goals undistracted. The Corporation can focus 100% on Firefox, and the Foundation can look after the project as a whole, including all the non-Firefox software (from NSS to Rhino via Seamonkey and Thunderbird), and also (in my view) run a wide-ranging donations program across the free software ecosystem. So let’s disentangle the people, even disentangle the technology and resource (maybe the Foundation finally needs its own sysadmins?) and let people focus. Then either option 2 or option 3 allows Thunderbird to take flight.

13 thoughts on “Thunderbird Thoughts

  1. That’s the first comment I read that brings the focus back upon “MoFo is not FiCo”. Many thanks for that!

  2. Interesting post Gerv. There is a difference between saying “Mozilla is not Google Labs and shouldn’t be haring off into 30 different new projects” and saying that existing, mature projects such as Thunderbird should be tossed. Let’s not forget that when Seamonkey was “communitied” one of the points made at the time was “it’s not like we’re getting out of the mail space – there’s always TB…”

    With the mess that email verification has become, it would be good to have Firefox pushing the notion of universal standards the way it has with the web and your reminder to the boosters of this idea that the internet is not just the web is very welcome. I saw one poster say “well FF has 50-100m downloads and TB only 5-10m so…” – that probably makes TB the second biggest open source app out there!

    Apple is already gunning for the non-IE browser space but it doesn’t have a Mail program to match Safari in the Windows environment. When it does, some mofos at moco might have cause to regret the fact that they couldn’t walk (FF) and chew gum (TB) at the same time.

  3. Gerv, one quick question regarding your last couple of sentences. How would it make sense to “disentagle” the hosting and IT folks support for pretty much everything — the shared version control systems, other shared apps like despot, bonsai, tinderbox servers and clients, the shared build and release equipment, the shared distribution networks and machines, AMO, AUS, Bugzilla, http://www.m.o, devmo, the wikis, community servers, etc.

    These are clearly all mission critical for the Mozilla Corporation so I don’t think that we’d move just move them out of our system over to the Foundation.

    Would the Foundation, which employs no sysadmins, no web application developers, no build system engineers, and has no experience in any of these areas hire up staff to develop and manage the duplication of that entire set of infrastructure for the not-Firefox Mozilla projects?

    That’s seems pretty wasteful to me and I’m not hearing enough complaints of poor IT and hosting support for the not-Firefox Mozilla projects (though maybe I’m not listening closely enough? Are you hearing things from the NSS or Seamonkey people?) to warrant something as wasteful as that.

    – A

  4. Asa: I’m not suggesting an enormous duplication of all the infrastructure. As you say, clearly that would be ridiculous. But I think it would be useful for there to be people on the sysadmin team whose focus was on non-Firefox projects (and who would therefore probably be paid by the Foundation).

    If you ask the SeaMonkey team, I believe one of their major releases was delayed several weeks due to lack of sysadmin time to do the release. I’m sure you would be hopping mad if a Firefox release had this sort of delay, so I’m sure you can understand how frustrating that was for them.

    But it was also a point made in anticipation of Thunderbird not being a MoCo product any more. If that happens, then clearly if we want to continue pushing Thunderbird releases, the resource to do that has to come from somewhere. My thoughts are but one idea; how would you manage it?

  5. Gerv, I doubt it was sysadmin time that delayed a release. Much more likely it was the overloaded build and release team. Sysadmins don’t really manage releases as far as I know.

    Yes, build and release human resources are tight and more of them would be good. Developing or attracting build and release people is something that all of the Mozilla projects need to think about.

    How would I manage it? A lot like it’s being managed today. Infrastructure and services that are critical to Firefox and other Mozilla-hosted projects are managed by Mozilla Corporation’s IT team and Web Apps team.

    As for build and release, for applications that utilize Mozilla’s AUS — which is Firefox and Thunderbird today and should include the other apps at some point in the future, we should have AUS setups for each Mozilla application, maintained by the IT and Web Apps teams. We should leave it up to the projects to manage their own build and release process, though. That would mean that Mozilla Corporation’s build and release team would manage Firefox releases and the SeaMonkey, Calendar, Camino, and Thunderbird projects would need their own build and release teams.

    That’s not to say that the Mozilla Corporation B&R teams wouldn’t assist when they can or provide advice and expertise where appropriate, but just as the not-Firefox projects have their own developers, testers, support, and promotion teams, they should also have their own build and release people.

    I’m pretty sure that this is already the direction for projects like SeaMonkey, Camino, and Calendar.

    – A

  6. Reed points me to this bug where you can see this arrangement in action. Mozilla Corporation has agreed to fund the setup, hosting, and maintenance of an AUS server for Calendar. Mozilla Corporation’s IT team will set up a VM on an existing box and will support that box and VM. The Mozilla Corporation’s build team is offering advice about the setup and how it can be used to also support SeaMonkey. This will empower the Calendar team to take control of their own build and release process and lessen the burden on Mozilla Corporation’s Build and Release team.

    – A

  7. Asa: You are right; I was being loose with my words. I was including “build and release” in my “sysadmin” category, both at the beginning and in the comments. It’s all BOFHing to me ;-)

    So when I said “disentangle the sysadmining”, what I probably meant was “have a separate build and release team”. So perhaps we are in agreement :-)

  8. I saw one poster say “well FF has 50-100m downloads and TB only 5-10m so…” – that probably makes TB the second biggest open source app out there!

    TB is also the second biggest open source app used for reading mail, the biggest one being Fx.

  9. For an outsider, the situation seems pretty clear. During the last few years, tens of millions of dollars have been flowing into the coffers of mofo/moco, almost all of them originating from Google (which BTW owes mofo/moco billions if you ask me, just calculate how much of their revenue is due to the fact that FF has google as the default search; but nevermind that). Anyway, Google’s interest is in developing the browser but not developing the mail client, which is a competitor of Gmail, and in which it is effectively impossible to implant ads. As, obviously, the main sponsor of an organization tends to call the shots, or at least influence the managerial echelons enough to shape long-term policies, Seamonkey and Thunderbird getting the ax is not something unexpected.

    But this is on the level of interests and power relations. On the more technical level, clearly the state of Thunderbird can be improved fantastically with more funding. KaiRo says there are 150 FF (or mostly-FF) developers and just 2 TB developers. And your and Mitchell’s talk about the ‘managerial overhead’ is just evasion. You’ve got overhead? Hire a TB development group manager (or if you want to be less hierarchical, an administration and coordination assistant). See? Problem solved. Mind you that this is completely orthogonal to the question of how exactly moco/mofo is organized internally (divisions, subsidiaries, for-profit, non-profit, whatever).

    What bothers me the most (ok, not the most; the googlebucks themselves are what bothers me the most) is the lack of intellectual honesty and the rolling of eyes. How can you write with a straight face that you “see no reason the Mozilla Project should not provide” more money for TB development? Especially when you “understand the desire to have an organisation fully focussed on Firefox”? Why can’t you call the “advance the open web” slogan what it is, which is a rewriting of the history of Mozilla to justify policies serving foreign commercial interests?

  10. Hum .. you already have some SIP support in the CVS repository, It’s called the ZAP branch (see for more information).

    I agree on the idea of having a some staff working on non Firefox projects paid by MoFo – that is a great idea.

    I also like the fact that you state, MoFo is not MoCo and that the Internet is not only web and web apps. This means that removing support for things like gopher (that was proposed in the last month) – should not happen.

  11. I think this brings up an interesting question: When the corporation was created, the Foundation was said to have 100% ownership. Now there’s almost the impression that the corporation controls the foundation.

    Seems to me like a fix would be for the foundation to fund a few B&R engineers to work along side those at the corporation, but focus on Thunderbird and non Firefox projects (is xulrunner in need of a home?).

  12. Nice post, Gerv. You’ve said pretty much everything I was hoping to hear. :) I’d like to see Thunderbird take Option 2 as you describe it. Imho it seems to be the better option both for Thunderbird and for the Mozilla Project and Foundation. If it’s decided a company setup more similar to Option 3 than MozCorp is desired for now, that’s still possible under Option 2’s structure, and wouldn’t relinquish the additional benefit of MF’s backing where wanted/needed.