Big day, right? iPad, Prop 8 Trial’s Witness Testimony concludes, Obama’s first State of the Union, Ted Haggard was cured of being gay (again). But you found the time to stop by my blog anyway. That warms my heart. So have a seat, and let’s talk about the add-ons manager.
Oxygen or Bling?
First, humor me a second, because I want to back out a little. Lately I’ve noticed that we talk about add-ons in very different ways. Add-ons developers and community sometimes talk about add-ons as something “everyone” should install. I’ve heard some say that there’s an add-on for everyone, or that everyone should install a few that help their browsing experience. This is the “add-ons are oxygen” view. And heck, it’s been said plenty that add-ons are one of the main things that make Firefox great, and they are certainly what differentiates Firefox from other browsers.
But hold on. Isn’t Firefox sexy and beautiful right out of the box? Don’t about half of Firefox users have no add-ons installed? And when users have problems with Firefox, aren’t “weird add-ons” the first thing we blame and disable? This is the “add-ons are bling” view.
So, should we encourage users who don’t use add-ons to install them? Are they oxygen or bling? Like about everything, they’re probably in between.
A Browser with a Bike Rack
The purpose of a car is to drive places. That’s why you’d buy one, and you’d expect it to do that perfectly as soon you buy it. There are some car nuts (I know quite a few) who care deeply about a car’s engine, performance, style, etc. They will make incredibly deliberate decisions about the car they buy, and even make modifications to get it to behave differently. Some of these modifications will be purely functional (e.g. tires with better traction), others are stylistic (e.g. racing stripes), and some help them perform other activities (e.g. a bike rack).
But the majority of car drivers won’t look for modifications and will be content as long as their car looks good, drives well, and get them from A to B safely. If they’re driving a good car, it’s probably because one of their car-head friends told them it was a good model (explains my Civic). Should these users be encouraged to modify their car? If so, how?
While the casual car drivers may not understand much about the workings of their car and may be bored at what car-heads find interesting, many could have a better car experience if some modifications were easy to understand and easy to install. If I could change the color of my car every day with a click, I would. Ditto if I could add a laser gun to the front. If putting a red stop-sign sticker on my windshield protected it from bird droppings, I’d stick it on.
So if you were a car manufacturer of free cars and free car modifications, and your goal was happy users, what would you do? First, for those car-heads, you’d want modifications to be easy to develop and share. For those casual drivers, you’d want the free modifications to be simple to find, install, remove, and understand. You’d want it to be clear how a modification would change the driving experience. And if one of those casual users wanted to become more of a car-head, you’d pave the path for them. These are roughly the goals of the add-ons manager redesign.
Add-ons as Preferences
There’s something else to keep in mind with the add-ons manager. Add-ons are essentially choices the user has made about how to tailor their browsing experience in a way that reflects their preferences. Uhoh, preferences? You bet – installing and configuring add-ons is a small part of a user’s browser preferences, and as such should probably not be separate from the Preferences menu. I constantly see users go to the Add-ons Manager when they want their Preferences, and vice versa. So in a future Firefox, it makes sense to combine these. But, because there are many separate challenges involved, for now we are focusing only on the add-on manager’s redesign, with the eventual goal of integrating it with Preferences. I’ve described in a previous post that it makes sense for the Add-ons Manager to display in the content area of Firefox: I think the same can be said for the Preferences menu.
The Preferences project is still in its infancy, but keeping in mind that the add-ons manager will probably exist within the context of the user’s Preferences is important both in design and implementation.
Design Direction
The design currently being considered for the add-ons manager is a two-panel basic hierarchy view within the content area of the browser. Add-on categories would display in the left panel, with an expanded on the right. This gives the user an overall view of that the add-ons manager entails, as well as enough room to individually configure items.
One of the benefits of having this two-panel design is that it allows scanning the contents of a category. The current add-on manager currently employs a sort of “digest-style” summary for each add-on, and with category view this sort of functionality can be improved with better information.
If a user selected a particular add-on from their manager, the right panel would show a detailed view of that add-on. This is something the current add-ons manager lacks: enough space to find out more about an add-on and configure it. Below an add-on’s information, the developer could add preferences for that add-on. The current manager pops preferences out in a separate window, launched from a button. This is essentially a pop-up window within a pop-up window. A two-panel in-content layout allows all of this functionality to be integrated.
This post is getting long pretty fast, so I’ll close here for now. Next up – how some of the interactions for maintaining, configuring, and installing add-ons could work within such a design.
As usual, these designs are not final and feedback (especially from add-on developers) is absolutely welcomed..
This looks great!
A few notes: Themes and Personas are both big categories, and both have a lot of active users, so I would expect to have each their own entry, instead of mixed together (as they are now). Also, with Personas, they can do instant preview and apply, but Themes (as of yet require a restart or re-opening all windows (see the ‘switchthemes’ extension).
Alfred Kayser –
Thanks for the note – figuring out how Personas and Themes live together is what I’m working on currently. The two do work differently, and users interact with them differently, but making this clear without being overly complicated will be a challenge.
This looks great, though I still don’t feel convinced that “everything is just a tab” is right. That said, the interactions themselves can live _anywhere_ regardless of what box surrounds it. A well-placed window with the same content area size achieves many of the same goals around size/navigation/not obscuring effects of changes, without dealing with the various visualization issues. Tabs on top can mitigate this, if the “browser” chrome isn’t shown, but bookmarklets etc all scare the crap out of me in these mockups.
mconnor –
Tabs are a bit cheating, because they can be torn off and treated like a separate (maybe chromeless) window. Removing that Chrome and making this more like a “home tab” with limited Chrome might work. Hell, just pressing back and not going up a level might throw people.
Well, I think it is certainly an improvement on what is there currently, but I worry if the categorisation will be successful enough for people to work out what is where. What’s the difference between a plugin, an extension and an add-on? Will people know which is which?
Logically, personas and themes might go together, but as Alfred’s comment points out, they (currently) behave pretty differently and have different implications for the user.
The mock-ups look nice, but my Firefox actually mostly lists various Java and Microsoft bits along with my anti-virus software, which lack icons and in some cases descriptions, and they can’t be uninstalled or rated.
Maybe I’m jumping too far ahead, but I worry that giving more space to the content from the add-on means that stuff that may be messy and confusing will get more space than it does now.
Your car analogy is interesting, because Firefox’s development model is essentially “Everyone gets Model A”. The add-on system ultimately fails for most users because the add-ons come without a warranty. They are stuck with Model A.
From this perspective, any solution involving a better search UI does not help users. What users need is a way to get more models with warranty coverage. They want 2 door or 4 door; value pack or sport pack. They want choice, but choice limited to option sets that work together. Better search should be “pick one of these 8 models, then select among 8 options on each model”. And all 64 results should be known to work.
johnjbarton –
Add-on collections probably go closer to creating that, where you can in a sense have a developer environment, or a social networking environment, etc. Still limited, and perhaps the performance drawback of installing multiple add-ons is not made clear.
I agree, this analogy is interesting, thanks for sharing
@mconnor regardless of what the choice is for Firefox we will have to be able to display the redesigned add-ons manager in its own window since for some applications the tab option isn’t there or relevant. It will be a per-application choice where they put it.
“66% of Firefox users have no add-ons installed” isn’t quite right. The post explains that at least 33% of users have add-ons installed, but we know it’s quite a bit more than that.
Justin Scott –
Thanks very much for the note – I changed it to about half to reflect the 53% of added blocklist pings added to the ratio to the add-on pings.
Correct me if I’m wrong, but wasn’t the OLD way of doing Addon preferences for them to make their own XUL panel and jam that in with the rest of the browser preferences? And then Firefox came along and they realized “hey, I need to configure Addon X, so let me go to where Addon X lives.” And I like that model very much.
Also, you’re not really distinguishing between chrome addons and content addons here. I have both. My chrome addons obviously extend the functionality, like Read it Later, Xmarks, and (for the moment), the Strata suite. My content addons are to enhance the sites I use (ie. Gmail) or to protect me (NoScript). Of all my addons, the only one I tell users is critical is NoScript. The browser, as it stands, works just fine for many people without any Addons.
alanjstr –
Going to where the add-ons live to modify them is very much the model to follow. The question is more where they live, and if they are findable where they live. Because add-ons are arguably more important than any other Preference item, as they have a full community and create virtually any change to the browser, it would be my goal to not obscure them in Preferences, but rather make them more findable as the result of any move.
It’s a good point you make about content vs chrome add-ons. We do not distinguish between them today, but perhaps that distinction would make sense in a future design. I can think of a few add-ons that are not fully one or the other, but finding meaningful categories that make sense to tech-saavy and not-so-tech-saavy users will be a main part of this design effort.
What URL are you choosing for the add-ons manager? In your screenshots the Awesomebar doesn’t have a valid URL.
I think the URL is important because this is the web; it means people can refer to the page, hyperlink to it, even perhaps call it with Ajax as a web service from an extension. Google agrees; Chrome chooses chrome://extensions.
Chris, I’m not sure the URL is terribly important to be honest. The majority of users will access this through the menus. That said we can easily choose anything we like.
We probably won’t be allowing people to link to it from the web for security reasons, and I don’t understand what AJAX services you’d expect this to expose to extensions.
Two things worth noting:
– Songbird does the Addons-as-preferences model ATM.
– Chrome does the second type of thing, where downloads/history/extensions are their own tabs instead of separate dialogs. I like this.
I do like the extended details for each addon shown in the mockup, and the in-line preferences.
OMG I’m a ricer.
The transition to a single addon seems odd in that you lose sight of the other addons; I wonder if a Finder-style multi-column view complete with fancy sliding would help. But then that gets rid of the top level selection, which sucks too…
The chrome-in-tabbrowser approach (complete with orange, err, Larry-area) is interesting; like AJAXy things like Gmail, I would expect the back button to go back to the previous selection, and _not_ the last web page. This may seem odd in the end, though – needs some playing around to be sure.
Does the “search in all addons” box search in installed addons (extensions/themes/plugins), installed addons in this category (just extensions), or on a web service (AMO)? I could certainly see random extensions wanting to extend that, so I hope the design has a hookable API.
The categories may be hard to deal with – for example, there are extensions in the Appearance category on AMO (not themes/personas); would one find, say, Colorful Tabs there? Choosing the right icon for Languages and Search Engines would be really, really hard, too, if you want to be impartial and things like that.
I wonder how scalable the contents would have to be (in some cases it might get constrained to small windows, which would be bad).
Having everything in a tab is seductive, eh?
I just wrote a blog post on that after reading the entry here – http://home.kairo.at/blog/2010-01/everything_is_a_tab_seems_seductive
Thanks for following the same road!
I say go for it on both fronts; the sooner the better! They would fit in nicely and get rid of the disparity, add consistency and simplicity all at the same time. A welcomed change to the new UI redesign!!
Please don’t make the Add-Ons window a tab. It shouldn’t be a tab because of the issues some of us raised in your previous posts. This includes:
– You won’t be able to see the effects of updating options on the web pages beneath the window any more. You’ll have to tear off the tab into a separate window, which the majority of users aren’t familiar with.
– There’s a risk that it could be spoofable. Who’s going to notice the absence of a differently-coloured logo in the address bar?
– It’s going to be easy to lose if you have a lot of browser tabs open at the same time.
– You’re mixing concerns. Setting preferences is not the same as web browsing.
– Browsers have traditionally had separate windows for setting preferences.
Please reconsider and keep this a separate window.
Looks good. However, for the average user, I think one of the biggest improvements would be making the terms “add-on”, “extension”, and “plugin” meaningful. These terms have distinct meanings in Mozilla/Firefox which are completely lost on the average user and, if anything, become synonyms, which is wrong.
One comment above has already suggested using the URL “chrome://extensions” like the Google browser does. But that would be totally wrong as extensions are just one type of add-on in Firefox speak. Just a case in point.
Using the car analogy: people know what a bike rack is, they know what hub caps are, they know what paint jobs are, and they would understand that these all fall under what might be called “customizations”. “Add-ons”, “Extensions”, and “Plug-ins” in common English all have the SAME MEANING and so there is lots of confusion. It would be very interesting to see how these words have been translated into other languages. It would be neat to see a post about that.
Great idea.
It’s about time for that change.
Are you planing to integrade jetpack extensions somehow in this new addon manager?
Iraê – Absolutely, integrating these will be essential
As preferences and Extension are the things I am most often accessing, I have a button for each and use FireFox with the MenuBar Hidden.
I like the idea of one preference to rule them all but how will this fit in with FF4.0 changes removing the MenuBar and replacing it with a App Button?
AppButton at left and PreButton at right?
:`’ I am very thankful to this topic because it really gives useful information “”.
Сергеевка-однин из интересных курортов Черного моря, где можно провести свой оздоровительный, недорогой отдых на Черном море. Главное, что привлекает курортниковна море в Сергеевку – это доступность в отношении цены, а качество не уступает 3-х звёздночным отелям.
Сергеевка-однин из курортов Черного моря, где можно отлично и недорого отдохнуть летом. Привлекает курортниковна море в Сергеевку – соотношение цены, и качества обслуживания. От нас уезжают чтобы вернуться.
Сергеевка-курорт Черного моря, где можно неплохо и недорого отдохнуть летом.Тихий, уютный городок с безлюдными черноморскими пляжами.Чистое черное море и золотой песок.
Pretty nice post. I just stumbled upon your blog and wished to say that I’ve really enjoyed surfing around your blog posts. In any case I will be subscribing to your feed and I hope you write again soon!