I think it's worth coming at this problem from the other direction as well, crushing it, finally, in something we like to call the pincing crab-claw of good design. Implicit in the question of "how do we help people manage tabs?" is the assumption that people want to or should have to manage their tabs; this is definitely true for some tabs, but I think that fully half of the solution will come through helping people avoid a situation that requires capital-M Management in the first place.

Reduce

Sometimes people have multiple tabs quite simply because they need to have multiple documents or applications open simultaneously. But sometimes tabs are stand-ins for activities that could be supported (better, even) in some other way.

• Applications
  A number of my long-running tabs are applications I refer to over the course of a day: webmail, a calendar, an RSS aggregator, an IM client. And the list of web applications that people use, often in place of traditional "web 0.0" apps, is definitely not getting any shorter. This is bad news from a tab management perspective — it suggests that our lists of perma-tabs will only get longer.

  One way to deal with this this is by getting web apps out of the browser and into their own independent windows, as in Prism. By handing the window navigation task from the tab strip to the OS, we let users make use of the often richer and certainly more familiar set of window navigation tools available there (alt/cmd-tab, expose, taskbars/docks, and so on). Even working almost entirely within a browser as I do, I still find myself quite often hitting Cmd-tab (on the mac) to get back to my email, which is, of course, just in another tab rather than another application. This doesn't make as much sense for transient documents, but for something long running like an application, this can be very powerful.

• An attention queue
  People often use tabs as a kind of to-do list — pages to read, forms to fill, reminders of tasks to carry out. Why as open tabs rather than bookmarks? In part, I think it's because an open tab is less permanent than a bookmark. Especially before Firefox 3, bookmarking a page felt like a commitment and required filing — not exactly right for something you just intend to deal with As Soon As Possible and may need no long-term connection to. It's also because a bookmarked page, while around for later, doesn't have the presence and judging stare of an open tab. An open tab can be an unbidden reminder in a way that a bookmark, hidden until called forth, cannot.

  What can we do to support the "deal with this later" use-case that doesn't contribute to tab-clutter? I've started tagging pages to come back to with the tag "queue" and then, periodically, looking at everything with that tag. It's not an ideal solution, though, because I still have to initiate the reminder process. It helps to remind of what I have to do, when I think to check, but not that I have to do something.
  

Re-use

Quite often, when I try to sift through and clean up my tab pile, I find that I have the same document open in multiple tabs. In part, this is because it's getting easier to navigate to a page again than to find one that I know is already open (which better tab navigation will start to reverse). But in many cases, I'm sure I just didn't remember that I'd opened something previously. Making it easier for people to re-use an already open tab would certainly help prevent the tragedy of tab-inundation.

One early concept for this is an attempt to intercede when we see tab duplication beginning — as a user begins to tell the browser where to go. Some quick idea sketching led me to some variants on putting this information in the awesomebar - more discussion of these in a future post:

(click for larger version)

Recycle

Lastly, some open tabs are documents that you're well and truly done with, but that you didn't close. Metaphorically speaking, you just left it on the floor and walked away, again. Would it kill you to tidy up after yourself?

Amidst the sea of open tabs, though, it takes effort to comb through your tab list for the flotsam and jetsam. To paraphrase something I read somewhere recently, there's immediate benefit to opening a tab but not for closing one. The benefit in the latter case is separated in time from the effort to get it — this is a contributing factor for all clutter, really.

What can we do to encourage or make it easier for people to close tabs that they're finished with? Some add-ons try to lower the effort bar on the triage, by providing an indication of tab age and disuse, for example. What else can we do here?

If you have thoughts or suggestions about these issues, please do get in touch! Other good sources for insightful thinking about the nature of tabs and their management are Aza and Bryan. Update: Andy Edmonds has also written on this topic.

which leads, inevitably and in the spirit of adventure, to this:

Now to find the right recipe...

Others of the changes are more subtle, and have more to do with simply lessening the effort involved in using the Manager. As always, improving the fit between human and software requires an understanding of what people are really trying to do — in the case of the Download Manager, people are usually involved in one of two main activities: (1) watching, and occasionally controlling, the progress of their downloads; and (2), finding files (or information about those files) that they've downloaded in the past.

Is it done yet?

In many ways, the best progress indicator is one that you don't actually have to watch. When a quick download finishes, it's useful to see the manager so that you can deal with the file right away. If a download is going to take a while though, people typically won't sit and watch — it makes more sense to get the Download Manager out of the way and get on with something else. It's still useful to have a cue that something is going on and an access point to the Download Manager, though, which is where the download progress indicator in the status bar comes from:

I remember downloading it...

The latter of the two main uses is that of finding files you'd downloaded earlier. To help out with this, we've surfaced details that correspond to what you might remember about the file you're seeking. Much of this information was previously available but obscured — you could see it once you'd found the file you wanted, but not when hunting for the file in the first place. For example:

• The time or date, in a person-friendly format, is listed down the right-hand side of the window for easy scanability. If what you know is roughly when you downloaded the file in question, you can run down that column until you get to the range that's relevant.
• The file name is obviously an important cue as to what a given file is — it's been made a little bigger so as to stand out amongst the other backup information we're now providing
• When the file name is vague (document1.pdf?) or not meant for humans in the first place (xpz243ghb.exe), it can be very helpful to know where the file came from. To that end, we show the name of the site from which you got the file. For me, this is a bit of help that I now find I can't do without.

Once you've found the file you're after, you can open it (double click) or remove it (hit 'delete'), or, if you're after more information, right-click:

Of these options, the one that's new and that I find particularly useful is "Go to Download Page." If what you're after is not the file itself but where you got it, this will take you right there.

But I know what it's called!

Scanning through the list is made more productive now, but searching is even better. Here's an example of how searching, combined with the information we now surface, makes finding a downloaded file much simpler. Say I'm looking for take-out menus — I can just filter on menu:

What's interesting is that few of the filenames themselves contain the name of the restaurant, but I can see which one I'm after because we've listed the site names. These site names (along with all the text you can see on the screen) are searchable as well, which saves me in this situation: I'm looking for a menu from Juice For Life, but their name isn't in the file name:

And there's the file. Happy searching!

I typed "new document" into the awesomebar, and there, in first place, was what looked like a promising link. I selected it, and got the following:

In other words, I was delivered directly to my real goal — not navigating to the front page of the Google Docs site, but to the "new document" I was trying to create. What's interesting, and newly possible through the awesomebar, is that I could find the appropriate link. Clearly I'd been there before, but I never would have remembered this kind of interstitial non-manually-typed-in URL. It helps, of course, that the URL/command in question uses actual human-language words.

As a result, I have functionality similar to a "New Document" button in my browser without having Yet Another Button on the screen, and the browser becomes an even more sophisticated intermediary between me and my online life than it was before.

Incidentally, Dria's written a great overview of the Awesomebar in Firefox 3. If you're interested in what else you can get out of this new feature, go read it!

I think the key to the theme of what we're working on for 3.2.1 and in 3.4 emerges from Morgamic's list of AMO's varied roles:

• A place where new users try to find add-ons to improve their browsing experience
• A hub for advanced users to pick up on the latest and greatest add-ons
• An incubator for new features
• A place where developers can get feedback and statistics for their add-on
• A tool we use to help QA popular extensions and ensure they meet quality and security guidelines

In the 3.2 redesign, our goal was to improve the experience for users in the first scenario on that list — people new to the idea of add-ons as well as the large and growing set of people who want to stop in, quickly find something to improve their online experience, and be on their way again. This constituency has, I think, been under-served in the past, which is the reason for their priority here and also for the coincident additions in Firefox 3 (the recommendations in the Get Add-ons tab, for example, should help to get new users started and oriented and also drive some new traffic to AMO). It's in this light that we made a number of the AMO changes, such as the more prominent search field, a deliberately trimmed down set of initial details about an add-on, the grouping of add-ons by purpose rather than implementation type, and by having the install buttons stop you before you've installed an add-on that won't work. My sense is that we've made positive strides for these users, though it will be some time before we can tell to what degree — figuring out ways to measure this with an audience we won't hear from directly is one of the tasks ahead.

All of that said, AMO exists equally for its other four listed roles. A successful AMO must, axiomatically, support add-on developers and those advanced users who are at the forefront of add-on testing and reviewing. It's worth noting, as well, that these sets of needs are not entirely orthogonal; in the same way that AMO being easy for casual users helps add-on developers by multiplying the latter's ability to reach a large audience, a site that's efficient for advanced users and developers is one that is more likely to be host to many great add-ons for casual users. More directly, in the redesign, making the add-on rating interface more prominent and listing experimental add-ons to everyone are changes that benefit both groups. The trick is addressing the needs of these two (or more!) very different audiences in ways such that neither's experience is diminished.

Basil's post included a summary of the top concerns we're seeing, especially from advanced add-on users and developers. 3.2.1 and 3.4 will see many of these addressed, especially in cases where we've made their use-cases less straightforward by hiding information that they find necessary but that confuses or distracts new users.

• In some cases, information that was pruned is making its way back in, placed to be, hopefully, findable by those who want it, but not overbearing to those who need not be concerned about it. Like putting compatibility ranges and access to all versions (for all OSes) back in:
  

  Or adding more visible access to feeds of new and featured add-ons:

  

• In other cases, we can address particular audiences by taking advantage of a logged-in state. For example, we can let advanced users disregard safeguards that don't really apply to them (the link enables the button):
  

• We're also refining categorization, including undoing some category merging that just didn't end up working, as in the Themes & Appearance category. There will be a category called "Themes" once again that will be solely and entirely about themes.

And, alongside all of this, Fligtar's amazing work on improving the usability of AMO's developer tools continues as well.

You can check out the preview site as we work on these issues — everything we change changes there first. Beyond this, too, please do keep the feedback coming, both in terms of the recent redesign, but also about how the site can better serve advanced users and developers. In some cases, there may have been ways of doing things on AMO that became more difficult after the redesign, but where the previous method hadn't been ideal either. Are there things you're trying to do that could be more straightforward?

Basil and Morgamic have already blogged about it, and what's new, but I wanted to pipe up and add my congratulations to the AMO/webdev team, AMO editors, and localizers for the really tremendous effort that's gone into overhauling the site.

This release is particularly exciting for me, partly because the wireframes and interaction-model for the site were the first things I worked on when I arrived at Mozilla. Also, though, because alongside the re-skinning and reorganization of the site, I think we've managed to improve the user-experience of finding and installing add-ons in a number of interesting ways:

• Being smarter about version compatibility
  In the past, a user has had to check to see whether the current version of his wanted add-on was compatible with the version of the browser he was running. I suspect that most users don't really know what version of the browser they're running (especially in a world with auto-updating) - nor should they have to! Comparing numbers is one of those tasks for which we have computers in the first place.

  We've built some logic for this into the new "Add to Firefox" buttons — we detect what you've come to AMO with, and present you with more human-friendly options:

  

  
  

• Making search more prominent
  
  A majority, or very nearly, of people looking for add-ons will search for them. In the past, our search field and controls have been hard to find. There's no missing it now:
  

  
  

• Making it easier to see preview images
  For a user browsing through the add-ons catalog, there's really nothing that delivers quicker understanding of what an add-on does than screenshots. A preview image is now a prominent part an add-on's listing, wherever it shows up; thumbnails of an add-on's preview images are visible at a glance on its details page; and users can bring up full size versions of each thumbnail (and advance through a slideshow) without ever navigating to another page.
  
  

There's more that's new, of course, and Basil's post does a great job of itemizing them.

Between the new AMO, the recent additions to the Add-ons Manager, and cleaning up the install process (a clearer install confirmation dialog and doing away with the install whitelist), I think that it will never have been easier to customize Firefox with add-ons.

Alongside that work, though, has been quite an effort to iron out some of the usability wrinkles in the process of installing add-ons. Some of these fixes have been in Firefox since beta 3:

• Simplifying and clarifying the message of the install confirmation "trust decision" dialog
  
• Making it clearer when a restart is required, and providing a convenient restart button at that point
  
• Automatically enabling the last-installed theme rather than requiring users to restart twice
  
• Eliminating the complicated "manage your whitelist" dance
  

But one of the worst ones on my list has remained as yet unironed: that Firefox abandons users after the restart required by add-ons installation. In the Firefox 2 task flow, a user interacts with the browser in a reasonable back-and-forth until he or she is required to restart the browser. After that point, the user is left hanging — the browser comes back up, but it doesn't offer any visual indication of whether the user has actually accomplished anything. In some cases, the newly installed add-on itself offers up a preferences window or steers the browser to a welcome webpage, but this is by no means universal. Regardless, it's the responsibility of the browser to complete this transaction with the user and to ease him or her back into the restart-interrupted train of thought.

As of today's nightly build, thanks to Dave Townsend and a merry band of reviewers working through Bug 408115, the Add-ons Manager will present itself, post-restart, to confirm that installations have taken place and will, helpfully, point out what's new:

Users can launch into setting preferences, start working with their new add-ons, or just get on with browsing, but, in all cases, they won't be left hunting in the interface for some sign of what just happened.

As usual, I humbly request your feedback — please tell me what's what here in comments or over in the discussion.

Compatible recommendations in Beta 3!

Laura's written an interesting post about her work on the AMO API, in which she talks about the its current state and where it's heading. One thing I'd like to point out is that some of the performance optimizations she mentions — for recommendations and search results — should be going in any day now. More performance improvements, related specifically to search, are to follow soon. Optimization is on the march!

Please leave comments here, or, better, join the discussion over in dev.apps.firefox.

Thanks to a lot of hard work from Dave Townsend, Laura Thomson, and the webdev team, you can, as of the latest nightly build, now explore add-ons from within the Add-ons Manager.

When you open it, you'll find yourself presented with a few recommended add-ons. You may see fewer than this in the near term, given that the manager will only recommend add-ons that are compatible with your version of Firefox, but more and more are compatible every day!

You can also search through the full addons.mozilla.org (AMO) catalog and install directly from the Manager:

When you install, you'll encounter a simpler confirmation dialog than in the past:

and, finally, we make it clear that a restart is required and give you a way to do it:

On top of these changes, the manager will now enable a Theme when you install it, so, one restart later, you'll see it in place.

This is a work in progress, of course, and there are some things still to come:

• having the add-ons manager be open, where you left it, after you restart Firefox to give you some feedback about what's happened
  
• highlighting your newly installed add-ons, post-restart
  
• continuing UI polish

Please try it out and tell us what you think, either in comments here or in bugzilla (Firefox::Extension Manager). We're eager to hear what you think!

If you're one of the few people who read this blog for, well, any other reason -- be not afraid. I'll still be using this blog for photos and blowing your minds, and all the Mozilla-related content will be tagged as such ("mozilla") for easy identification.

Shh shh. It'll be OK.

• Man Throws Red Paint Into Trevi Fountain (AP)
• Dye in the Trevi: Some Romans See Red, but Others Cry 'Art!' (New York Times)
• And so on

So, in the public interest, the service of which is, after all, the primary purpose of this blog, here are some photos. You can also link to them right in the Flickr.

Sorry. Back on topic. I went down to Harbourfront on Sunday afternoon to take some pictures. I'm quite happy with this set, never mind the high swan content.

I've also been trying out a new flash-based photo display system (yes, another) called SimpleViewer. Have a look at this set in SimpleViewer and let me know what you think.