An update on Thunderbird’s support plan, Get Satisfaction and SUMO

Since my last blog post about a position being available for support, I’m pleased to announce (belatedly) that we’ve hired Roland Tanglao to lead Mozilla Messaging support. It’s not obvious being the only person on staff supporting millions of users around the world!  In order to learn how Thunderbird support is currently happening, he’s been busy immersing himself in that world.

Success for anything user-facing on the scale of Mozilla Thunderbird requires scalable approaches.  One person, no matter how well qualified and efficient, can’t help everyone who may have issues (especially as many people come to us with questions that are really about their mail provider).  So we need to look to systems and communities to help us help users.

Get Satisfaction

The first system that we’ve committed to is Get Satisfaction, an online website which is designed from the ground up for peer-to-peer support.  You go there to ask a question, and maybe you also help someone else who has a question you know the answer to.  The “take a penny, leave a penny” model.

There are two things that stand out for me with Get Satisfaction: their user experience, and their APIs:

Get Satisfaction clearly thought long and hard about what the user experience of a support site should be like, and it shows: Threads are categorized as being ideas, problems, and praise, and problems and idea threads can be marked appropriately when either there’s an solid answer of some sort, or to track the evolution of ideas.  This provides people who want to dive into a support role with great dashboards, which makes dealing with a high volume much easier.  It also provides users a way to communicate their input & mood effectively, which is an important part of a support interaction.

The second aspect of Get Satisfaction that was particularly compelling to me was that they have built their system with HTTP APIs in mind from the beginning.  This means that we can integrate it into our own websites, or even in future versions of Thunderbird, without anyone needing to rework the database.  We’re just starting to figure out how to use these APIs, but I’m hopeful it will allow us to streamline support interactions considerably.

You can find out more about Roland’s thoughts about Get Satisfaction in his blog post.

Knowledge Base

Even as we’re happy with the capabilities of Get Satisfaction, it’s clear we need to complement it with a knowledge base where we can build longer FAQs, helpful documents, etc. There, we got an unexpected assist from the Fennec (Mobile Firefox) project, as their need for a variant of SUMO for Mobile users made it so that the SUMO project had to refactor and “productize” SUMO.  Having more than one installation of SUMO supported by the core SUMO team made me feel that our use cases were likely to be supported well, which is great — we try to leverage as much of the Firefox infrastructure as we can, as efficiently as we can.  Thanks Fennec!  (For now, we’ll stick with Get Satisfaction, and won’t be using the SUMO forums or the Live Chat features).

SUMO will allow us to build a scalable, localizable document store — it’s been proven to scale (handling 22 million hits a day for Firefox!), and has a built-in localization system.  Given that Thunderbird 3 will likely be published in 40 languages, ignoring localization isn’t a realistic option.  (I should point out that while GetSatisfaction is currently only in English, that team is keen to explore internationalization and localization).

For English users, MozillaZine is the obvious alternative to both Get Satisfaction and SUMO.  And indeed MozillaZine currently has lots of great content, and a bunch of people who provide great support.  Unfortunately, it also lacks a bunch of features (APIs and RSS feeds), has an interface which we can’t improve upon, and is English-only.  An interesting observation is that with people relying on search engines ever more, it’s not really a problem to have multiple sources of information about a product.  I expect MozillaZine will continue to thrive, and we’ll definitely continue to work with the all existing knowledge bases on the web and link to them appropriately.

A web of communities

Going forward one of Roland’s tasks is to build or strengthen the bridges between Mozilla Messaging and the excellent Thunderbird related communities on the web. Mozillazine, of course, but also geckozone in France, and as many other support communities we can find.  Around the time of Thunderbird 3 launch (hopefully in November), Roland is planning some sort of worldwide support community online meeting. If you’re interested in participating, contact Roland.

One of my hopes is that better organization of support communities can lead not just to happier users, but to a better product.  One of Roland’s duties is to provide feedback to the product development team on frequent or emerging issues faced by users, or areas of Thunderbird that particularly please people.  That’s not something he can summarize accurately by himself — he’ll need your help.

Positive Feedback

Support is too often a one-way medium, with users complaining about problems.  That’s certainly an important function, but in a cooperative open source project, it’s equally critical that QA and developer contributors thank users for their input, and that users point out parts of the experience that they like, as positive feedback is the best motivator.  So take some time out and pick an open source project and generate some good karma!  For Thunderbird, it’s easy, go to getsatisfaction and click on “give praise”.  Roland will make sure that the people deserving of your praise get to see it.

Support job with Mozilla Messaging

frustration, from e-magic on Flickr
"frustration", from e-magic on Flickr

One of my first jobs in IT was as a “computer consultant” for my university.  I got to learn a lot about computers of various kinds (including currently useless but still formative bits like writing REXX programs on a CP/CMS IBM 3090 mainframe), and, more importantly, I learned a lot about what it takes to really help people solve computer problems.  We had all the kinds of users you’d expect in a decent-sized community: haughty faculty (and haughtier grad students, for some reason), who absolutely demanded that you stop whatever you were doing to help them fix their margins <em>right now</em>.  We had to mediate between first-year students lost and confused in their first exposure to the net, and technical staff who were straining under the onslaught of the first week of classes.

Still, it was a lot of fun — we had autonomy to decide what problems were worthy of paper handouts, to hand out as answers to FAQs — we had special accounts with free access to the high-resolution printers — but most of all, we had lots of interactions daily with people who were truly grateful for the help we provided.  Not all interactions were positive, but the vast majority of them were.  There was a real community between the student staff, the full-time staff, and the avid users who spent their days in the computer center working on their papers, homework, and assignments.

The job

These memories come back as we’re now looking for someone to help coordinate Thunderbird’s technical support communities.  The challenge is orders of magnitude harder, but the opportunities to help are equally huge.

Mountain Lion Safety by ekai on flickr
Mountain Lion Safety by ekai on flickr (helpful, no?)

It’s become clear to me that this job is fairly unique, both in scope, and in what we’re looking for.  The right candidate will have a rare blend of empathy & technical knowledge, clear organization skills.  She or he must have the ability to bring people off the ledge, recognize and encourage peer leaders, but also know how to deal with poisonous people.  This requires both clarity of thought and clear, efficient expression.  Our ideal candidate knows Thunderbird well, but most importantly can understand both the requirements of providing technical support for hugely diverse populations of users, who often come for help when they have critical issues that are often caused by external actors like email providers.  A hard job to fill.

A big part of the job is also to understand the existing bits of the existing world of Thunderbird support, and figure out ways of connecting the bits that work, and building new bits if needed.  From where I’m standing, this ecosystem includes forums like mozillazine, which have garnered huge numbers of posts with very valuable information, but which don’t necessarily provide the best experience for novices.  It also includes the GetSatisfaction forums, which provide a lower barrier-to-entry for many users, but which are still in their infancy, both in terms of content and in terms of people providing answers.  The ecosystem also importantly includes SUMO, which right now is focused on Firefox, but which could hopefully be deployed for Thunderbird use someday.  A comprehensive solution also likely involves figuring out how to help people over Twitter, on Facebook, or wherever Thunderbird users are likely to look for answers or express frustrations.  Whoever gets this job will end up diving deep into the world of Mozilla support overall, and learn from all of the people who’ve built those foundations — then build upon them.

As the central point of contact for support issues, this person will be incredibly helpful in working with the QA and dev teams to prioritize bugs and other issues, so that the next revisions of Thunderbird or Thunderbird websites work better for more people.  Finally, it would be great to coordinate with the support micro-communities that exist around the world.  In general, the job is clear: help people with Thunderbird, and help people help each other.

If this sounds like you, send us an email, explaining why you’re the right person for the job.

Getting insight into one’s own email

Thunderbird knows a lot about your email. After all, it’s got access to large amounts of data, and builds sophisticated databases so that it can be very responsive, even when dealing with large folders containing thousands of messages. Wouldn’t it be nice if we could use this information to extract either faster workflow, or even insight into our email habits?

On Tuesday, as he was slogging through some somewhat antique code in the front-end of Thunderbird, Andrew asked whether Thunderbird’s “start page” was what we wanted it to be, which is Andrew-speak for “this is broken, we should fix it”. We’d actually discussed what to do with that page for a long time, but had focused on other, more infrastructure bits for a while. I decided it was time to dust off those plans, and see what we could do.

First, some background. The “start page”, which makes a lot of sense in Firefox, never made a huge amount of sense to me in Thunderbird. In particular, it’s shown only when a folder is selected, and no message is selected. That’s hardly a logical time to show the (colorful, pretty, but fairly useless) page we show now. Instead, why not show information about the selected folder, and help people who clearly intended to select a folder, so most likely wanted to do something related to that folder!

We could still use a small part of the pane to display interesting news snippets, as a way of keeping users aware of new developments, such as new add-ons that we want to promote, or Mozilla events. In fact, by making that page useful, we’re more likely to get people to read a sentence or two which might change periodically, as opposed to changing a page that is so big, wordy and useless, that I suspect most users are completely blind to it.

A couple of days later, we have an early patch, which feels pretty compelling to me:

  • It summarizes critical data about the folder (name, number of messages, number of unread messages);
  • if you were recently reading that folder, it points out the last message you were reading, so you can quickly find it again;
  • it shows you a few of the messages that are most likely of interest (because they’re unread or starred);
  • and finally it gives some information about the activity in the folder: a histogram showing activity over the last 52 weeks, and most prolific authors and most active threads during that same period.

For example:

Folder summary

The icing on the cake is that if you hover over an author or a thread title, it will highlight when those messages occurred in the histogram (in this case, the “An alternate take on HTML 5” thread). Readers of m.d.platform won’t be surprised at any of the names that show up above!

There are lots of possible additions: Some of the ones we’ve thought of include:

  • showing how much disk space is used by the folder
  • showing the “largest” messages
  • showing the status of autosync/indexing jobs

Note: this hasn’t been through a visual design phase yet — the look will change, once someone who knows what they’re doing has a chance to make it nicer! There’s also a lot of work to do to make sure that it does the right thing for saved searches, smart folders, etc. And yes, we’ll make it optional, just like the start page is now.

I’ve been playing with it for only a little bit, and it’s fascinating how interesting it is to get some data on one’s email. For example, I more-or-less follow a small-inbox model, where emails don’t stay in my inbox if I can avoid it — they get deleted, or archived. As a result, the sparkline for my inbox is:

inbox sparkline

which isn’t as short as it should be, telling me pretty quickly that I have some old messages that need attention. It’ll be interesting to see what else we learn from these kinds of simple personal analytics.

What other kinds of visualizations, summaries, and analysis would you like to see in Thunderbird, or in add-ons?

[Update]: I liked the idea that a commenter suggested of showing what tags were in use in the folder, so I had some fun, first with a “tagpie”, which proved mostly useless, just like most piecharts; then with a “tagdots” small-multiples visualization:

tagpies, tagdots, sparklines oh my!

I don’t think the tagpie is worth keeping, but the dots are tempting. As an aside, the Protovis library really is as easy to use as I’d hoped.

Thunderbird-friendly SOGo reached 1.0

An open-source project that I pay a bit of attention to is SOGo, aka “scalable OGo”. It’s a project w/ a long history, much of which I don’t know well, but what’s interesting to me is that they’re trying to be the Yin to Thunderbird’s Yang. SOGo includes a web frontend, which is designed to look like Thunderbird, which is a fascinating choice.

The folks at Inverse, who are the primary drivers of SOGo, as far as I know, are also doing Thunderbird & Lightning add-ons (SOGo Connector and SOGo Integrator), including adding support for CardDAV, Free/Busy, etc. They’re also helping with patches to Lightning, for which we’re definitely grateful!

SOGo reached the 1.0 milestone yesterday, which is always worth celebrating. If you’re interested in the groupware space and looking at software that plays well w/ thunderbird, check it out.

Thunderbird 3 beta 2

On the road to Thunderbird 3, another milestone — this time, Thunderbird 3 beta 2.


Why do beta releases?

Beta releases are funny things. They serve a few purposes. The first is to make sure that we periodically stabilize the code base, as without periodic ‘cooling’, it’s hard to get a handle on the quality of a piece of software. Betas also serve as deadlines, which are magical motivators for some people. Some of us will spend way too many hours staying up late in the night in order to “make a deadline”.

Betas for open source software are even more odd in that people interested in staying very involved with the project can use nightly builds, which are updated every day. I’ve been using nightly builds of Thunderbird for over a year, as have several thousand other users. As a user of nightlies as much as a project coordinator, by the time the beta is released to a wider audience, all the excitement is historical.

Another fascinating aspect of beta releases is that, because we know there will be another release, and because the purpose of the beta is to get a broader set of testers to shake out edge cases, we try to be conservative about slipping in major features at the last minute, as the odds of those features being polished in time are never what we hope they’ll be. So we routinely delay feature additions until the next cycle, to avoid dragging the beta validation process out. It’s an unpleasant, but unavoidable part of optimizing releases.

If we do our job right there, then by the time the beta ships, the features that have landed are free of major bugs. We of course can’t know that until we get feedback from the beta.

What’s in this release?

The most striking part of the release is the sheer volume of bug fixes. It’s not sexy work, it’s often the hardest work, but it’s very important. This list (of bug fixes and feature work, but mostly bug fixes) is impressive.

Of the features that have landed, I want to talk about two that many users could easily ignore: archiving, and the activity manager.

The archive feature is straightforwardly borrowed from GMail’s archive feature, which we think is great. The idea is that figuring out exactly which folder each message should be filed is a process that can take a lot of time and effort — something that wasn’t a real problem in the early days of email, but which becomes a real time sink with thousands of messages. With a good enough search engine, it’s easier for many users to simply “archive” the message (doesn’t really matter where), get it out of the way, and then rely on the search capability to find the message again.

In this beta, we’re half-way there. The archive feature is there if you want it, but you can also use the standard “file in a folder” method. Thanks to work we did before beta2, the archiving is fast, putting messages in per-month folders at the click of a buttton or a keystroke. The new fast global search hasn’t landed yet, but even our “old” cross-folder search mechanism has gotten a lot better.

I already love the feature — being able to select messages I don’t need to worry about anymore, hit ‘A’ and be done with them, saves me a lot of time and mental effort

The second feature worth highlighting is also not fully deployed, but already useful. The Activity Manager was born out of a recognition that Thunderbird 2 is pretty bad at telling you what it’s doing. It says a lot of things, it says them fairly loudly, but they’re rarely the things you want to know. We’re building infrastructure that will let the various bits of Thunderbird be much more helpful in describing what’s going on (through a log of notable events), what went wrong (non-intrusive but notable alerts), and how it’s progressing at long-running tasks (with more context than just a single progress bar). Teaching software that wasn’t designed with a notification mechanism or philosophy in mind how to be polite and informative is a slow and arduous task, but we’re making good progress. In Thunderbird 3b2, there’s an Activity Manager window, which for now will just report on message moves, copies and deletes, and IMAP auto-syncing. Now that the framework is in place, we should be able to have a lot more informative messages when you need them, and reduce the number of dialog boxes (especially the ones you can’t do anything about!).

One of the fascinating aspects of the activity manager is that it’s giving even those of us who know how the software works on a detailed level a better handle on important global aspects. For example, the activity manager showed me that the autosync function can and should be much more aggressive, so that more of your email is already downloaded before you need it.

Other features you may notice:

  • Much more useful Growl notifications on OS X
  • Keyboard shortcuts for quick tab navigation
  • Better looking forwarded mail
  • Fewer dialog boxes

What’s next?

The next beta release is our last scheduled beta. As such, we’re thinking of it as the last milestone to introduce Big New Features. Furthermore, we’re hoping to be even better behaved this cycle and land features as early in the process as possible. Upcoming features which we hope will be available in a nightly build soon include:

  • the new global search function, leveraging tabs
  • cleaning up the message header area further
  • “pop tarts” to complement the activity manager
  • the beginning of some theming work (prettier icons, etc.)

And then, of course, there will be unplanned bright ideas which show up out of nowhere. Life wouldn’t be fun without those.

Try out the beta, file bugs, send feedback!

PS: the illustration at the top is from a brand spankin’ new website for Mozilla Messaging. We’ve changed the site to make it the primary destination for Thunderbird users, riffing on the look of other Mozilla websites, and yet quite distinct. I find the illustrations in particular a lot of fun, and I’m very proud of the team that built it. Rafael Ebron ran the project with the SpreadThunderbird team, with designs from The Royal Order, and implementation from silver orange. A very nice job, thanks to all who contributed!  The new site also allows us to build localized sites, which will be amazing.

Lightning-in-Thunderbird status update

For some time, we (the Thunderbird release-drivers) have been exploring how to best integrate calendar functionality into Thunderbird.  Time for an update.

The current plan is to work with the Lightning Add-on community to make a version of it available as an add-on to Thunderbird 3 after we ship later this year.  This is a change from our initial plan of integrating all of Lightning into Thunderbird by default.  Our thinking has evolved based on both technical and product reasons:

  • the calendar team has had a tough time of figuring out how Lightning needed to change to integrate optimally in Thunderbird, in large part because Thunderbird 3 itself has been somewhat of a moving target — Thunderbird’s interface model is still shifting, which is certainly hard on add-on developers
  • partially as a result, there’s a fair amount of work left to do on Lightning before it’s ready to present as part of the core product.  For example, it’s not yet possible to selectively enable the task management or calendaring features; the account configuration is a bit too hard to get right for novice users; error messages are still too cryptic — that sort of thing.
  • looking at Lightning as an addition to the core project, it would represent a lot of new, complex code, with a lot of implied maintenance — we don’t have enough developers on hand to be able to take on that commitment at this stage.  See Philipp’s post on this topic for more information, especially if you’re keen to help out.
  • we try to avoid feature changes in between major releases, and limit minor updates to security fixes.  This would mean that feature changes to Lightning would have to wait for the next major release of Thunderbird.  That would be a real shame.
  • more generally, we recognize that different users need different kinds of calendaring solutions.  Just as there are more and more messaging systems, there is a growing diversity in calendar usage models, such as web calendars, stand alone clients, and calendar and event applications on social networking platforms.

Given all this, we feel the best plan is to take advantage of our add-on ecosystem, to allow a variety of calendaring features to evolve, and to allow Lightning to evolve at whatever pace is best for it, less tied to Thunderbird’s schedule.  Lightning is by far the most popular and important Thunderbird add-on, and we’re going to see what we can do to make it better and more successful, both on the development side, as well as on the promotion side.  Also, thanks to Thunderbird 3’s new add-on manager, it will be easier for users to find and install, which I expect will lead to even greater adoption.

In short: expect a version of Lightning that will work with Thunderbird 3; expect it to have more frequent releases than Thunderbird major release numbers; talk to philipp if you want to help!