Office 12 and the Task-Based Interface
The Change
Today Microsoft has given the world a sneak peek at Office 12. Given that there’s a fundamental paradigm shift in the interface and a lot of strong reactions, I figured now would be a good time to briefly introduce task-based interfaces to those unaware of them..
I present to you the topic of contention:
These new tab.. menu.. toolbar.. things are somewhat of a new way of interacting with Office. The idea is that Microsoft is going to use a more “task-based” form of interaction, allowing the user to select the kind of task they which to perform and go from there. It’s a laudable goal.
On Being Task-Based
In general, the user interfaces we use on our desktop and laptop computers today are application-based—they are entirely centered around the notion of applications. You start Microsoft Word and edit Microsoft Word Documents from within it. You can then switch to TextEdit and edit the same Microsoft Word Document if you so chose. They’re two separate applications, yet both can work on the same kinds of documents. The Macintosh interface, originally, was helping to push some document-centric notions into the interface. All of a sudden your documents were nicely presented in folders and they remembered information about themselves. You double-click on a document and it opens up in the last application that saved it, rather than whichever application is the default for that file type. This was part of why the Macintosh interface did not need file extensions—the file type didn’t matter as much as the application which saved it. That added document-centric notions, but the interface was (and is) still extremely application-based.
The problem with application-based interfaces in general is that they are “modal”—you must pick which application you wish to interact with, and are confined to the actions presented by this application. There have been many attempts to modularize pieces of applications so that this is less of a problem, but none have been ultimately successful. Jef Raskin writes:
The present structure of computer software, consisting of an operating system under which application programs execute, is inherently modal. This implies that for an interface to be nonmodal, an approach that does not include applications in their present form is required.
Jef Raskin – The Human Interface – 5-7 Applications Abolished, p. 139
Eliminating everything that we have done for the last twenty-five years is not easy, so no one has shown much interest in completely removing applications in favor of some currently-undefined muddle. But that doesn’t mean that interface designers can’t learn from task-based thinking.
Game interface designers certainly have learned—Sim City 3000 (pictured right) is just one of many games to use a more task-oriented control center. From the top, the five big buttons represent the major categories of task: manipulating the land, zoning, constructing transportation, connecting utilities, and placing government buildings. The controls are organized by what the user (or player in this circumstance) will be doing, rather than by which controls are most similar from some other feature perspective. For example, Sim City 2000 clustered the controls for zoning near the controls for constructing government buildings because they had to do with creating buildings, except for the airport and seaport zones, which were clustered nearer to the utilities.
Microsoft’s interface designers have shown interest in task-based interfaces for the past several years. The redesign of the Control Panel in Windows XP was intended to make it more task-based. A lot of the new “wizards” in Windows XP were obviously along the “task-based” line. Help systems in Windows XP and other modern graphical interfaces have also obviously been informed by task-based thinking instead of the traditional feature-based thinking.
So, realistically, it was just a matter of time before Microsoft’s task-based experimentation led them to rework Office. Going from a traditional menu/toolbar interface to a form of task-based interaction is difficult, especially when there are nearly two decades of tradition built up in our current widgets (I use the term “widget” to refer to what are often called “controls” or “interactors”—they’re the buttons you press, the menubars you click, the scrollbars you scroll).
Do or Do Not; There is no Try
Widgets, actually, turn out to be my major gripe with this redesign. Microsoft is completely reinventing the Office user’s workflow. I cannot emphasize that enough. This is unlike any way these users have typically interacted with an application. Yet what we have is this:
This combination of widgets look a lot like things we’ve seen before, and not very much like something new and special. I have yet to come across an actual name for this interface. People have been calling it all sorts of things, but there is no centralized marketing name for it. There’s no big push from Microsoft to call it something new and explain exactly what it is as if it’s something new. It feels more like Microsoft is putting it out there as something that’s “sort of new”. It looks like old interfaces, but doesn’t really act like them. It acts like tabs, except it doesn’t switch the document view. It acts like menus, except not really because it acts like tabs. It shouldn’t really look like either.
The visual design is what really gets me. There’s the typical Vista-esque overuse of gradient. There is a kind of subtle gradient like Apple tends to use in the backgrounds of its newer applications, yet Microsoft wants to use it to differentiate between controls as well. It’s like they’re saying “this is a nice background effect… but it’s also important for telling things apart.” Pick one or the other, don’t overuse the gradients for both. It becomes quite confusing.
In general there seem to be some issues with space. They are still trying to cram too much into too little space and still end up with something which looks cramped and messy. It’s difficult to know how exactly some of these smaller 12″ laptop screens are going to fit all of these icons—I suspect that’s why there’s a lack of spacing between widgets.
The buttons are jammed together and the grouping is not very clear (in part due to lack of whitespace). It’s difficult to tell, exactly, which piece of the interface has control over what without actually trying it first. It certainly needs to have more of Don Norman’s infamous pseudo-affordances (or virtual affordances). How exactly is “Write” different from “File” other than a very slight baseline shift and a very subtle background color shift? There are arrows all over the place—if you have a group of common widgets with common functions, the layout should probably not be based on the length of the label like that. We don’t place the right-facing arrow for normal submenus next to the label, we place it at the side of the menu, so why should we place these downward-facing arrows next to the label here, other than the lack of enough differentiating whitespace?
If you click the image I presented you’ll get an even larger one. In that you will see both the words “Insert” and “Contoso Legal” presented in some sort of tab-like form. It may just be my screen, but it’s awfully difficult for me to differentiate between the two—yet one is an extremely important control over the entire display of the lower-half of that screenshot!
If we’ll return to my Sim City 3000 reference above, you’ll notice that their toolbar has an extremely clear mapping between the large “task” buttons and the smaller “subtask” buttons. There’s a nice colored background linking them all, and they are not cluttered together. It is extremely clear what one can and cannot click on. The buttons which require further user input even have a tiny little white square with dots in it, indicating a dialog (similar to “…” in menus today). Maxis goes one step further and even color-codes the task groups. When you click the “manipulate environment” task (as I call it) at the top, the background color grouping becomes a nice solid green. It’s very clear which task you are currently in and there is little chance of confusion.
Summation
I think it’s wonderful that interface designers are starting to seriously look at task-based interfaces, but I do not think they are something to be taken lightly. If Microsoft wants to be serious about it, they should make it a big deal. They need a name for this interaction point in Office. They need a clear branding of it, and a clear visual design. Unfortunately, “clear visual design” does not seem to be the current direction for Vista’s interface. Less muddy gradients, more solid differentiation of widgets, please!
This direction is interesting, though, and fortunately this is not a released version of Office yet. The final version certainly has the opportunity to become something much better. Whether or not the rest of Windows Vista will manage to be equally compelling… we’ll see.
Update: I see that the area I’m referring to is “code-named the ribbon”.
I personally like the look and feel of the beta they have just released. Much better than Office2003 blue gradients.
New users might enjoy the task-based UI. However, switching from feature/tab-based UI to tasked-based UI will require some learning curve. It would be really frustrating when you know what you want but you can’t find it. But I agree that this is a nice direction to go into.
How about a search-based office UI? Type in (or speak out) what you want and the application shows what it can do for you.
The colors and gradients and the like are certainly prettier than previous versions of Microsoft software, but it’s now much harder to identify what is interactive. When switching UI paradigms, the most important thing is getting the users to understand how to interact with the new UI. Making it harder to differentiate elements and figure out what is what is the exact opposite of what Office needs to be doing.
Yes, disrupting users’ workflows is a very serious thing, especially when you’re doing it for millions of users. That’s part of why I think Microsoft needs to be taking this whole thing more seriously—marketing name, extremely clear and unambiguous interface, etc. It’s either that or don’t do this. They can’t do it half-way and expect users to get it.
How is a search-based UI fundamentally different from a command-line? Even with Spotlight you need to know what you’re searching for. At my command-line I can begin typing and hit “tab” and it’ll search for matching command names. I agree that this kind of functionality is useful for power users, but we can’t pretend it’s something for everyone. The command-line is the ultimate in power user interfaces, and I expect that we’ll have more search-friendly and voice-friendly command-lines in the future—but they’re not a replacement for all of the real hard work on the regular interfaces.
I haven’t seen the MS presentations, but I can’t make head or tail of that screenshot. I see three horizontal layers of stuff beneath the title bar, and I don’t understand how they relate to each other. In particular, “Insert” and “Contoso” line up with each other, so are those part of one thing? My eye jumps at the border and sees “Insert / Contoso / Our Logo”. But I suspect that’s not the way it works at all.
Apple made a very serious effort to move to a document-based interface a decade ago with OpenDoc. You can’t fault us for not trying! I’m not sorry it was killed, even after I’d spent 3 1/2 years of my life on it, because it was a fatally flawed architecture and implementation. Still, no one has attempted anything similar since.
Jens: Exactly. :)
interesting entry. at last MS risks at something that matters (UI). I cannot stand UI as is now and the “it’s old and tested and it works” is not good enough for innovation. Let’s see when OOo will catch up
I hope — whatever changes they make — MS makes it easy for me to put things back the way I want.
A little humor about the new UI.
Shiney Blog
Actually, however bad it may be, I hope that it is not optional. People who want the old UI should use the old version. If Microsoft’s UI becomes worse, customers need to complain about it, not just turn it off.
You either do it or you don’t, you don’t say “maybe” and add another option. MS Office applications have more than enough options already.
Actually, however bad it may be, I hope that it is not optional. People who want the old UI should use the old version. If Microsoft’s UI becomes worse, customers need to complain about it, not just turn it off.
It’s not that simple. Many users are in environments where they will be forced to use the new version whether they like it or not. And no matter how “well designed” the new interface is, there is going to be a learning curve associated with learning it, as well as a productivity drop until people become comfortable with the new UI. Who is going to absorb this cost? A lot of people may say the new UI is better, but is it really? The hubris of designers who are convinced they know best never ceases to amaze me.
If the new UI is not optional it makes my decision over the new version much easier: No.
http://mrshiney.froppy.com/blog/
If you’re not willing to accept the cost of upgrading, then don’t upgrade. That’s how it always is with new software. Making it optional is not actually going to change a whole lot of things—there are many reasons additional options are evil, including increased code complexity which leads to an increase in bugs, which in the end would lead to a higher cost anyway, especially when we’re talking about two entirely different user interfaces.
Seriously, I hope they make the new UI not optional, and I hope that if the cost is to high for you, you don’t upgrade. I’m in no position to care that Microsoft won’t get your dollars right away.
Some day, in the future, they may refine it to the point where it will be worth it for you to upgrade (especially if the majority of Vista is based on these new ribbon things and you eventually upgrade to Vista). Until then, feel free to not upgrade.
I’d rather see them get less money from people choosing to not upgrade and really get this stuff done correctly than have them maintain backwards-compatible options (which make the software worse) for people who don’t like change (Word’s white-text-on-blue-background anyone?).
In corporate environments, you shouldn’t be upgrading unless you understand the costs of doing so and are willing to live with them. Yes, learning new interfaces is a cost, and it can be a large one. I never said otherwise. There’s really no need for corporate environments to be blindly upgrading just because Microsoft releases a new version.
Honestly, this looks simply like they are grouping toolbars at the top, which is a good thing. They are also adding realtime editing so you can see what a selection will do before commiting to it and having to click undo. But without really working with it, I don’t know how different this is…
One question from a usability side, how much SLOWER will this system be than drop-down menus? It takes a lot more mental energy to find the right icon, in the right place, with the right label than using well categorized menus… right?
sloan, in theory this would be significantly faster once it has been learned. There are now gigantic icons for each of the common actions. You click the right category (like “Insert”) and then it’s just a matter of fast icon recognition rather than reading words. If you’re relatively familiar with the icons, and the icons are differentiated enough, icon recognition can be really fast.
Traditional menus are known to be quite slow to navigate. There have been many proposed replacements which attack different problems, but none have really taken over for various reasons.
“The visual design is what really gets me.”
Can’t remember if it’s mentioned in the Channel9 video, but I’m sure they said that the visual design is NOT the final design - its only function is to provide the test and usability teams with a working program.
Jenson Harris mentions that the visual design is not the final design in his blog:
http://blogs.msdn.com/jensenh/archive/2005/09/15/467956.aspx
Yes, Microsoft says that about all of their betas, and things sometimes do change significantly, but that doesn’t mean the current stuff isn’t bad. All that means is that there is some hope.
It’s somewhat silly to send something off to the “usability team” if the visual design is causing a lot of confusion.
I know that it will change, but I’m not sure that it will change significantly enough.
I haven’t tried the beta yet, but stumbled upon your article.
I am studying for my PHD, and found your article interesting… especially the comparison to a game-based interface. It reminded me of a good quote:
[Videogames] They are so far ahead of the race, compared to the dull and workmanlike interfaces of “serious” software”
[S. Poole in Trigger Happy: Videogames and the Entertainment Revolution. ].. your article illustrates this nicely!
if only more “serious” apps borrowed from the good design employed in many games..
It’s called Results-Oriented UI.
Jakob Nielsen just posted an Alertbox article called R.I.P. WYSIWYG on the new paradigm. http://www.useit.com/alertbox/wysiwyg.html
The reason Microsoft is calling it “Results-Oriented UI” is because it does violate some of its Task-Based roots, but it’s certainly not the death of WYSIWYG. Nielsen’s comments about it being not WYSIWYG are just… odd.
There was further discussion about this on peterme.com.
From the other articles I’ve read I’ve come to understand that the Ribbon is basically a task-based interface grafted onto a tool-based interface. This is not really a “new” paradigm. The half-and-half-ness of the interface is where a lot of my uneasiness came from, I think, and I still stand behind my comments wishing they would take this all the way. I think half-way isn’t going to get them where they want to go.
[…] As I went into detail bout before, I agree with John on that last issue. […]
[…] 目前Office 2003使用者界面的工具列型表，大体上沿续自1997年开始采用的模型，这个多数人熟知的使用界面，将随着下一代Office 12（代号）的推出，而有了命革命性的改变。① […]