Julie Larson-Green rocks. If you fancy yourself having some interest in HCI and interfaces or even in computers in general AND you do not know who is Julie Larson-Green you should be ashamed of yourself. She is the woman who re-invented Microsoft Office, and recently she has been tasked to do the same with the whole OS for the next version of Windows (which will be just Vista with some polish, exactly like Windows XP that was Windows 2000 with some extra polish, as reflected in the version string going from 5.0 to 5.1 — shall we call it Vista.1?). It remains to be seen how that will turn out, but i for one have high hopes.
Anyways, a newspaper from Sydney (Australia? Man, what are people doing in Australia? There’s nothing there, is it?) had a profile of Larson-Green, which is cool since she definitely deserves more limelight (she seems to dislike it, at least in that vid i saw once).
What i want to comment, though, is actually just one small phrase she spells out on the interview, which is a powerful “quiet move” in that it redeems the old oxymoron of user interface theory of “intuitive use” without compromising the perspective of making user’s lives easier. The phrase being:
We want to reduce the amount of thinking about the software that they have to do
The discussion usually goes that this or that software is “intuitive”. What that is supposed to mean is that a given person when using the thing for the first time does not need training. That she will just understand the thing without explanations and that she will have a competent behaviour from the beginning.
But there is no competent behaviour without training, just as there is no unskilled use of computers. Intuition has no effectiveness in dealing with computers.
In this kind of discussions, proverbially the newcomer to the computer has been called the “grandmother” which is not only demeaning, it is also extremely misleading — in the sense that it diminishes the understanding of the persons who derisively talk about how a given program is or is not usable by grandma. So for now i’ll talk about a Completely Unskilled Computer Operator — or C.U.C.O..
So, back to our argument, when a c.u.c.o. first sees a computer screen with little pictures of trash cans and sheets of paper, he does not have an “intuition” that those are “documents” and “deleting files drop icon”. If you ask him what the paper sheet icon is, he will probably answer in a relatively safe way, which might seem to confirm expectations that he is “understanding” what is in front of him, but let me ask you: what exactly he “understood”? Did he understand that said thing is a “file” which actually resides in a disk and needs to be “saved” before being “closed”?
He is supposed to have understood that said thing — whose nature and behaviour he has no hint about whatsoever — is “meant to be used as if if was a piece of paper”. In other words, you gave the user very few reliable information, overloaded his expectations about the thing by telling him it can do things that it actually can’t, and most of all you forced on him a given mode of behaviour: write here! You forced him to act in a given way and proceeded to call that “being intuitive”.
This becomes even more dramatic if you remember that files can and should be used in more than one way. Not only that, the possibility and ease of doing so are some of the most valuable characteristics of computers: they allow us to thinker.
Now the fallacy of “intuitiveness” has already been thoroughly discussed. Alan Cooper for one has a very good critique (though he proceeds to achieve some very silly conclusions based on it). Suffice it to say that this particular question has been around computers for a long long time — we want to make users’ lives (and even c.u.c.o.s’ lives) easier, and it seems that those fancy things like GUIs and mouses have done so, but no one seems to pinpoint exactly how…
What is brilliant in Larso-Green’s take is that yes, it wants to do things for the user, it wants to take him by the hand, but it is not deceiving itself about some mystical properties that it clearly does not have like “intuitiveness” or “familiarity”.
She focused on “reducing the amount of thinking”. In this way, she actually keeps the effort to make people’s lives easier through fancifulness without the burden of the whole hot air of “intuitiveness”.
I am currently not aware whether she actually sees the connection i am implying here. Maybe her statement did not have the intention i am ascribing to it. Either way, i believe that HCI research should focus on this better account of the benefits of a good interface: the capability to diminish the needed effort to operate a given software or machine.
When we think like this, some of the previous trade-offs of interface design become actually straightforward. Classic example: to use or not to use wizards? Now the answer depends on whether the wizards will diminish or augment the effort needed to use the program. If the user will only be annoyed by the sequence of “press next”, then no, but if said sequence will actually ease a difficult decision, then yes. The thing is: now we begin to see that, if the presence of a wizard actually diminishes the effort needed to use the program, probably the interface is overly crowded to begin with.
It is no surprise, then, that Larson-Green actually managed to make Office’s interface better without relying on any of the “dumbing down” that many account for the success of Microsoft Windows. Her interface is cleaner and fancier at the same time.
From a HCI-theory point of view, the striking thing is that she managed to make the interface “easier” without most of the ad-hoc “features” that “intuitiveness” advocates usually rely on. In her own words: “we wanted to make the user experience more predictable, with less guessing and auto features“. More predictability — seems almost like something you could find in Linus Torvald’s bashing of Gnome — and still we are on a “make the user life easier” mission!
To me, the best example of this stance is in the Friendly Interactive Shell’s design document: “Every configuration option in a program is a place where the program is too stupid to figure out for itself what the user really wants“. To figure what the user wants is a computational problem now, it is not a “configurable option”. It is not something outside the scope of the software.
Obviously, the program can’t divine every whim, but indeed we can rule out a big deal of ambiguity that makes the program more difficult without making it more useful. Part of the problem is to think of “user productivity” without thinking in user behaviour: to measure a software by what it can do outside of a context.
It should be also noted that in Larson-Green’s stance there is little or no special treatment for the c.u.c.o.: we certainly do not want to make his life harder, but there are also no expectations that a user will be able to divine what a program does without previous experience. Consistency and an environment where experimentation are encouraged, maybe, but no mystic guessing.
This kind of thinking, even if it seems like a “quiet move”, means a big and important development in interface theory. Let’s hope more of the software world will pick this soon.