Thoughts on Thunderbird’s Evolution

Aside from all of the organizational stuff I’m doing like recruiting (more news soon I hope), dealing with the incorporation papers, budgeting, etc., I’ve been spending a lot of time thinking about Thunderbird’s evolution.

Specifically, how should we, as a project, figure out what to work on?

Some of my thinking about this hasn’t changed since I joined, but many of the details have, in reaction to talking to people on IRC, on the newsgroup, and, significantly, by watching a lot of bugmail.

First, a reality check. Thunderbird exists, today, and works, today, for large numbers of users, large numbers of use cases, against a variety of servers, in a huge array of different environments, from individual home users to massive organizational deployments. That’s a precious asset which I don’t want to either ignore or disrupt. So for example, any plans for Thunderbird evolution which start with “first, start over”, or “first, remove feature X” aren’t likely to go far.

Furthermore, there is an incredible store of community knowledge about known issues, good features to add, known architectural challenges, etc., which I’m slowly absorbing. And there is existing momentum there, with bugs being worked on daily, which I most certainly don’t want to stop. In the shortest term, all I’m likely to suggest are small course changes, although they’ll all be motivated by bigger picture motivations.

As an example of that kind of small influence, a little story: Mark Banner has been working on the Address Book for a long time. Joshua Cranmer, a more recent contributor, started to help out with a long-standing bug to convert the address book store from the “idiosyncratic” Mork database to the shiny new SQLite database now available. My contribution to that effort was to suggest in IRC that it’d be good to start with some unit tests, as changing backends without a test suite struck me as risky. A few days later, Mark had the beginnings of a test suite ready, that Joshua can use to move faster with the conversion. The bigger theme there is clearly that I believe we should build enough infrastructure such as test suites and test farms so that we can take on architectural changes with more confidence. That’s a theme you can expect to hear again. Much of the work that MailCo can do in the short term will be to help with setting up a context which is favorable to product evolution, so that people outside of MailCo find working on Thunderbird as rewarding an experience as possible. There are a surprising number of people who have tried to contribute to Thunderbird over the years, but there was no one with resources and the overall vantage point to help them help the project. (A related effort which I’m hoping to launch soon is to make it much, much easier for people to build, distribute, and install Thunderbird Extensions.)

There’s more to leading a project than making work safer or easier, though. Many people on the project would be happy to help move the project forward, if they only had a clear picture of where the project was going. Knowing about that long-term direction also makes it easier to figure out which of the existing feature enhancement requests (numbering over 1300 in bugzilla alone) fit in with the long-term vision for the project.

There’s a clear need for a long term vision, from which we can build a roadmap, taking into consideration the platform roadmap in particular. From that document we should be able to specify release goals, and then prioritize work items, identify areas needing contributor involvement, etc.

My approach to identifying this long term vision has been two-pronged. There are macro-level, contextual drivers, and there are micro-level, operational constraints. Understanding the current state of the project, is crucial to identifying what’s doable with a small paid staff and a lot of code, users, bugs, etc. On the flip side, the societal and market trends help us figure out where it’s worth trying to take the project. For example, instant messaging use is now pervasive. Newsgroup usage is declining. All other things being equal (and they never are), I’m more likely to encourage IM integration than newsgroup tweaks. But it’s a balance, because we may have (as we do) existing code and contributors who care about NNTP (such as the aforementioned Joshua!), and I’m not about to stop them from fixing bugs!

Other high-level points which influence me and will influence the long-term vision:

  • Mozilla is unique in that among all of the “vendors” of messaging technology, it is the only organization driven by the public benefit. This should allow us to meet the user’s needs directly, without having to get distracted by exit strategies, analysts, etc. It also makes it easier to recruit volunteers!
  • Thunderbird has a unique position in the market due to its relationship with Firefox, the larger trends regarding open source adoption, and its technology base. The Mozilla platform’s extensibility capabilities are unparalleled, and this gives Thunderbird huge potential for fitting in “niches” that are hard for others to reach.
  • The world of messaging has changed radically in the last decade, and not always for the better, from the user’s perspective. With increasing volume, spam, and diversity in communication modes, interacting via computers is getting more stressful, not less. From a product management perspective, there is “customer pain,” which means that there are problems to be solved. That’s a good thing!

In some ways, then, I’m forming a picture in my head which has a faraway, slightly fuzzy picture, of the Thunderbird that could be. And I’m also learning a lot daily about the Thunderbird that is. Interpolation is then just a small matter of design, architecture, coding, testing, inspiration, perspiration, patches, reviews, tests, parties, t-shirts, arguments, celebrations, conflicts, and conversations — lots of conversations…


  1. I agree with most part of this.

    However, I really think that “blank piece of paper” is right path for Thunderbird. It doesn’t mean that everything has to be replaced, but rather that everything needs to be rethinked. If Thunderbird “as is” had potential, it would make it, and it just didn’t. Just because lot of people worked isn’t an argument, so many people worked on Pentium 4 but it turned out to be the end of the road.

    Also, I think that extensions are not right path to go. Extensions have some consumer potential if they are released by big sites. They might also have some potential for real big companies. Other than that, they are just nice instruments to keep community in a good mood, as everyone that loses his feature or favorite way of doing something can use extension. And they are nice as a source of idea to implement into the core package. But not more than that.


  2. There are a surprising number of people who have tried to contribute to Thunderbird over the years, but there was no one with resources and the overall vantage point to help them help the project.

    IMHO, that’s key to attracting quality volunteers. I’ve scaled back my contributions significantly — not that they were ever enormous — simply because it was never clear my efforts yielded many results. As examples, triaging and/or patching bugs that are never landed wastes my efforts. I’d rather invest my limited time in something that produces results, which requires more organized efforts.

    On another note, right now we’re porting a business from TB to Outlook (due to a compatibility requirement for a critical 3rd party app). Some users are near revolt over the responsiveness, performance and ease of use they are losing. While we can’t compete with Outlook in 3rd party app integration, let’s remember our strengths as we move forward.


  3. First: keep up the communication! I like this blog, it has been hard to find any news or updates on thunderbird development and plans in the past. the wiki and webpages are often quite old.

    – Thunderbird is already very useful, it’s rare for me to think ‘I wish it had’. However, with people getting so many emails, speed (especially on imap, moving between messages) is crucial.

    – My biggest problem is I have thunderbird at home, work and on another laptop. I *really* wish that I could share *filters*, addressbooks, and other settings between these installations. (yes I know ideally imap filters should be server side but often not practical)

    – My organisation used to use Mulberry. This had *many* faults, but the nice thing was that I could log in from any PC across the organisation, at home, at a colleagues house, and get all my settings. the window would even resize to how I normally have it set.

    – I and many people have large numbers of email, and folders. Filing emails seems to be never ending, and finding them a pain (though I do think thunderbird’s search is better than others). An idea would be to look at new ideas for helping with this. Auto suggesting folders? keyboard shortcuts to move to popular folders (eg F6 to move current msg to ‘finance’ folder)


  4. Good luck with this opportunity David. I am a keen user of Thunderbird and Firefox and would concur with your overall appraisal. Thunderbird works, very well… So don’t screw it up. I am a big user of NNTP for mailing list access via gmane so enhancements in this area would be very welcome. I am also starting to use the Lightning extension/plugin in anger. So please don’t break the extension/plugin facility. I am not sure what planet Ivan is on when he says:

    Also, I think that extensions are not right path to go.

    Why was/is Firefox so successful? I’d suggest the extension capability is a major factor…

    Good Luck again and Happy New Year.



  5. I disagree with Ivan. If someone wants to write a new mail client feel free but now that Mozilla’s relationship with Thunderbird has diminished (the least contentious word that represents my opinion) available resources must be directed towards evolution rather than revolution.


  6. I agree with Chris Keene.

    Sometimes I had the feeling to be the only Thunderbird users left on this planet. Keep us informed.

    During the holiday I used a dialup connection with my IMAP foldes. I felt that the synchronization of IMAP folders lacks of information: the user doesn’t know what’s going, if the syncronization is finished or is still going on. I ended up using TCPDUMP under Windows to snoop the port 143, but I don’t think that this is what many users do. 😉


  7. I agree with the comment about synching various settings between installations of Thunderbird. The same could be said for Firefox.

    Along the same line, I hate having contact information in various places (phone, gmail, facebook, linkedin, corporate ldap, etc) that are never in sync. This is very annoying.


  8. any plans for Thunderbird evolution which start with “first, start over”, or “first, remove feature X” aren’t likely to go far.
    At the same time, any plans that call for Thunderbird to try to be just an open-source, generic email client are doomed to also-ran status IMHO. But I’ve already stated my case in earlier comments to this blog, that you need to rally us around a Big Idea.
    Here’s another Big Idea: champion a major upgrade or replacement to the IMAP protocol that would allow us to store message database information in a central location. People are clamoring for shared access to their email from multiple locations, but many of the clever things we would like to do in email and communication classification and context linking will really only work with a powerful email database (equivalent to an Exchange datastore).
    That’s possible locally, but not possible to share with current IMAP limitations as I understand them.
    I agree with other comments that leadership, and effective communication from that leadership, are key. Anyone who has ever had to make tough product management decisions with a limited R&D budget knows that either someone makes tough calls for priority, or you will end up with an undifferentiated, uninteresting product.


  9. To echo some of the comments of others, Thank You for keeping us informed. One suggestion if I may is for account settings do like Outlook and have ALL the sever information entered for each account. While I don’t use Outlook, I walk people thru setting it up on a daily basis (gotta love tech support). People coming from Outlook are going to be a bit confused (I sure was) with the way Thunderbird has you setup the outgoing server. Just my 2 cents. Keep up the good work.


  10. Address book improvements are much needed:
    – birthday field
    – Photo
    – more flexibility adding custom fields
    – The address book should be able to associate sent and received e-mails, notes, appointments, chat logs with a person. (similar to the Activities Tab in the MS Outlook Adress books). This would be a contact-centered view of all
    interactions with one person. I would highly appreciate this feature.


  11. I agree with guanxi that it is discouraging and a waste of time to contribute patches that are never reviewed. At least it would be nice with a message saying “I don’t have time to review your patch right now but will get back to it in X weeks” or “You should request review from Y instead”.


  12. Why do email programs for example, have to create huge email inboxe files and when you accumulate really big (90,000+) email files do they have to crash so badly. face it, people accumulate huge amounts of emails today, these email clients must be able to handle/back-up/manipulate these files easily. Going to switch to thunderbird soon as netscape won’t be supported soon..


  13. You’re right in identifying the pervasiveness of IM in recent years. However I think the use of email from multiple devices such as laptop, mobile phone and desktop to name but a few is very important.

    With Gmail’s introduction of IMAP, it’s already something that is in demand.

    I think this should be something that Thunderbird should strive to achieve.


  14. I have been a Thunderbird fan for years, ever since Microsoft dropped the free of charge ability to read Hotmail from Outlook Express. Ever since then I’ve used Thunderbird to read my IMAP mail and webmail accounts. This is also were I believe Thunderbird is competitive and has an edge on other mail clients and even webmail. The ability to have all email in one application, using one address book, ability to read offline and inlne spell check. Great stuff.

    At the moment all this is available with help from some addon’s. But this meakes Thunderbird a poweruser’s tool. To be more competitive the setup of email accounts and especially webm,ail accounts must be must simpler. No user can figure out how to download, install and configure to addon’s to get Gmail or Hotmail to work. This must be made simpler with a wizard, which also donwload the appropriate addonss and configures them at the same time. If we could get Gmail or Hotmail or any other to support Thunderbird as prefered offline mail client, usage would go up dramatically.

    Small problem is that Thunderbird does not automatically detect online/offline and keeps asking for passwords etc when it cannot connect to webmail.

    Lighting is also looking better al the time. But no calender is any good if this cannot be shared by different users. In office environments there is off course Outlook, which is hard to beat as it is bundled with Windows/Office. So the best option here is to link up with OpenOffice which is still lacking this functionality. Fot home users Exchange capability is not required, just beaing able to share address books, calander etc via FTP etc. would be enough.

    Keep up the good work!


  15. The bigger theme there is clearly that I believe we should build enough infrastructure such as test suites and test farms so that we can take on architectural changes with more confidence.

    Other projects like Evolution and KMail may already have large test suites for email that you could reuse (and collaborate on). There’s no sense in each project recreating tests for the same protocols.


  16. First of all, I’m happy to see that Thunderbird is apparently in good hands.

    I agree that although TB, as it is, is far from perfect, it is the base we need to build on.

    I also agree with the need for a long term vision, and to me IM integration, better support for extensions and moe focus on the futher development of the Lightning extensions are certainly great ideas. While the use of e-mail may be declining amongst home users, it will remain vital for organisations. So developing TB to become a real alternative to Outlook is the way to go.

    Keep up the good work!


  17. PostPosted: Jan Wed 30th 2008 12:24pm
    I have several gripes about Thunderbird. I want to use it, I like it, when it works.
    1. There are no install instructions !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

    Downloading and Installing
    System Requirements

    Before installing, make sure your computer meets the system requirements.
    Downloading Thunderbird 2

    Mozilla provides Thunderbird 2 for Windows, Linux, and Mac OS X in a variety of languages. You can get the latest version of Thunderbird 2 here.

    For builds for other systems and languages not provided by, see the Contributed Builds section at the end of this document.
    Installing Thunderbird 2

    Please note that installing Thunderbird 2 will overwrite your existing installation of Thunderbird. You won’t lose any of your mail, but some of your extensions and other add-ons might not work until updates for them are made available

    help about about shows no update available.

    I can’t believe that no one has fixed the inline picture problem or the attachment problem from rich text emails.

    I can’t believe no one has made a shell integrating Sunbird to make an outlook clone/beater. You could choose to use tbird singly or with the combined shell.

    Firefox is awesome, Thunderbird could be too if you guys would get BASIC functionality working.

    I hate Outlook, but I use it because I’m sick of repeatedly trying to deal with attachments and image filled emails not working.

    Please get on the ball

    This is unacceptable. Let’s get Tbird to where it needs to be, SOON


  18. Two things that have recently bothered me:

    1. Calendar is required. Lightning isn’t ready for prime time as of 0.7. Even with the obvious fixes, however, it MUST link to Exchange for office use. I’m running a Linux desktop for everything except meetings, for which I must rdesktop to a Windows terminal server. It’s the ONLY thing I need the terminal server for. Then, however, I have to ALSO accept the meeting in Lightning so that I get a popup in my main window instead of a shrunken rdesktop window.

    2. Mailing list setup is less than pleasant. There’s no way to grab existing email addresses from the address book, it’s manual typing of each email address. Made my wife screaming mad the other week and when the wife is unhappy I’m unhappy. Please, for the sake of domestic tranquility, put some time into this feature!

    Otherwise I’ve been pretty happy with the application and look forward to using it for a long time.


  19. I think that TB has to move to an embedded database to manage email messages. With the exponential growth in passing multimedia files around via email the file/folder storage will be unworkable soon. Then again, if IMAP became the default email standard with 100 Mbit connections, the IMAP server could handle efficient storage. Or create an attachment database to store attachments after stripping them off the email message with primary keys stored the email message to retrieve attachments.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s