There’s a new personal e-zine, The Idea Basket, that’s debuting today with an editorial on the commercial feasibility of a next-generation user interface. The editorial talks about the state of current user interfaces, software such as Apple’s iPhoto that is pushing the envelope of interface design, reasons why there has recently been such a lack of innovation and research in the field of interface design, and some ideas that both consumers and developers can employ to help bring computer software to the next level of usability.
The Commercial Feasibility of a Next-Generation User Interface
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSAlert.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
… but iPhoto is pretty cool. Didn’t find it difficult to use. But then again, I used only 1.0. Maybe 1.1 changed the UI. I don’t know…
The problem is not so much how to make a neat interface for a simple
beginners’ program, but how to make a good interface for big
complicated programs like Lightwave or Photoshop or CuBase.
The idea that you could just drag-and-drop some “data” on one of these
is simplistic. So is the idea that they should all have matching
interfaces. Each addresses a different set of professional skills.
The interface of a computer is not just the appearance of the
Finder/Desktop screen. An important aspect is how clearly the OS is
structured. Can it be understood by the reasonably intelligent user?
You cannot show anything useful about an interface with screen shots.
Why are these interface articles always written by people who have
apparently never used Amigas, and in particular Directory
Opus/Magellan ? Indeed, some of the books take all their examples from
Windows.
Time to post this again:
http://research.microsoft.com/ui/TaskGallery/index.htm
This project seems to be a very feasible enhancement to the usual Windows-Icon-Mouse-Pointer (WIMP) ‘desktop’ norm.
…I haven’t heard of WIMP interfaces for a good 15 years!! Brilliant. For some reason, people started talking about GEE-U-EYEs around 14 years ago and then they started talking Gooeys and that’s how its remained ever since.
The first WIMP I remember was GEM……..
Frankly I don’t see any necessity for what the author proposes. If it ain’t broke, don’t fix it. Isn’t this really the case of a jaded consumer, perpetually dissatisfied with his new toy as soon as it’s obtained?
Jared is making his personal lack of satisfaction everybody else’s problem. That kind of thing is nothing new. The irony is that the only person who can make Jared happy is Jared. The failure of the god of consumerism is also nothing new. And the answer is not to try to reinvent the same old failed god. After all, the definition of insanity is to try the same thing over and over, expecting a different result next time. Jared needs to learn that true happiness will never come from buying things.
From a practical perspective, the “tools as entertainment” genre is terribly inefficient. It takes a lot of work to churn out an endless supply of cosmetic surprises. For those of us who actually use tools to be productive, the constant rearrangement to suit the passing fancies of others is a major annoyance. There’s nothing wrong with a little makeover, but the key is moderation. There’s a reasonable amount of change, that which brings true progress, and then there’s that which is just plain nuts. I’ll take quality over churn any day!
As for organization, well an organized person will be able to organize his life. If you don’t know what you want, no amount of programmatic cleverness is going to satisfy you. A fickle person is never satisfied by definition. Rather than try to make a computer program into an ersatz mommy, the middle-aged immature user should finally grow up. Jared, you need to be weaned.
I think Jared is right on. I think the idea about the key-word filtering is exactly what we need. Very interesting article.
Just a couple reasons (not all of them) why we haven’t we next generation UI’s:
The techies who design software (this goes especially for linux/open source developers) are extraordinarily hostile towards Human Computer Interaction people. They are seen as whiny bastards overly concerned about GUI stuff who tell programmers how to do their jobs. They are often told that if they want to change the way some interface works to make it better that they should shut up, put down their cognitive psychology books, stop that useless user testing, and pick up a compiler and code the better interface themselves. Sort of like a construction worker telling an architect that if he wants the building built durably and beautifully he should shove that useless drafting pencil up his butt and pick up a jackhammer.
The businesses that employ said techies really don’t see it as worth their while to make their products easier and more efficient to use. “But that’s crazy..”, you say. No, that’s not crazy, that’s the computer industry.
Businesses that buy software often think that one piece of software is just as good as another as long as it performs the task at hand. Considerations like cost and pure functionality (e.g. can it do thing A, thing B, read X file format) are considered more important than how efficient it makes the employee or how much less frustration they feel with the software at the end of the day. Cost and features are important, sure. But businesses can’t seem to understand that an interface that allows a secretary to write 12 letters a day instead of 8 and that leaves them feeling far less stressed is the most important feature they can possibly have and will be far more cost effective no matter how much less they could buy the other piece of software for.
Consumers think that computers are naturally hard to use and they generally don’t question this. They just say “I hate computers” and leave it at that. They’ve already been so terribly disgusted with their computers at work where they have to type out letters on a frustrating and inefficient word processor that when they go home they say they say “no thanks to all the other computer stuff, I’m just going to keep it to e-mail and web surfing, thank you very much”. And guess what happens if you have an economy that in a very short period of time has based itself almost entirely on people using technology to do far more than just e-mail and web surfing? How successful do you think a company that sells dog food over the internet would be if their target market was millions upon millions of people who say in unison “I hate computers”?
What goes around comes around.
why doesn’t some bored college programmer sit down and write somethin like this for linux (easier said then done). i mean how feasible is it for that to happen?
How did you all succeed in reading an article with such background and foreground. The first page gave me a headache.
I do a lot of work teaching adults how to use windowing systems. And I can say that the standard WIMP design is most definatly broke and need fixing!
Why?
Well simply put it takes ages to learn. It is most definatly not intuitive in this respect! And when you have learn’t it, it not very fast. I spend most of my time sat on a UNIX comand line, becase it is so much faster than the graphical interface to get jobs done.
The solution?
Not easy. My hunch is that a very radical redesign is required. More so than changing the arrangement of icons, menus and dialogs. More so than the rather naf, microsoft 3d gallery approach posted earlier… Anybody see anything truly redical?
Chris
As well as the other good things Ilan Volow said, I feel the need to again stress that the solution to disorganization is not a new system; the solution is to UNDERSTAND and then USE the systems you have.
WIMP works. Maybe there’s something better, or a refinement of WIMP. 3D is not the solution. More eye-candy is not the solution. Until the “something better” comes along, learn to use what you have, folks.
If you’re disorganized, that’s your own problem.
Haha, yeah… and this an article about UI, the UI on that particular site is just obsense, to say the least.
” the solution is to UNDERSTAND and then USE”
but arn’t computer suppose to help us and not us help them? they are suppose to make our lives easier by bending around our needs, wants, and preferences. that is the solution.
i applaud Jared for starting a new e-zine and hope he will continue to post thoughtful articles. I basically disagree with most of what he said. Everyhing goes back to the fact that the vast majority of people do not think abstractly. They think visually. Programmers think abstractly extremely well as do people who shine in math, etc. The WIMP interface is what made the computer revolution. And that’s because the vast majority of us could not and cannot “envision” the abstract of Unix, etc. The desktop interface is still it, it still makes sense and, most important of all, it makes sense to the average user.
Some of Jared’s examples are interesting. I don’t think he sees that, for the average user, programs like Kai Hagan’s are a nightmare. You go in and have no idea what to do next. That is the way things are.
Linux is an example of people who want an OS to chip away at Microsoft, but are not that eager to do what it really takes to grab the intermediate user, much less the averge user. Despite the progress that has been made in ease of installation and KDE and Gnome, Linux is a horrible desktop experience. People want to see icons of their drives, they want visual, not abstract. They need visual. That’s the only way most people can “get it”. And the desktop metaphor – that’s what still makes it work because people can get that. Once they get a feel for it, the light bulbs start turning on and how the whole thing works starts to make sense.
I grieve for the BeOS. It had the best of all worlds.
Ilan Volow, the easier solution is to teach kids how to use a computer.
Perhaps 9 years is a good age to start learning it.
> The problem is not so much how to make a neat interface
> for a simple beginners’ program, but how to make a good
> interface for big complicated programs like Lightwave or
> Photoshop or CuBase.
Another example on this topic is AutoCAD. It has a very complex user interface that maxes absoutley no sense to beginners. However once you have learnt to use it you realise that the design is no accident at all and it is specfically designed to let you put lines onto the screen as fast as possible.
I was just remarking at dinner about how in today’s society, computers let some people get away with behavior that nobody would allow if it didn’t have a computer attached. At work, people whine about how they can’t figure out how to use the computer.
But it’s part of the job description.
Any other job skill, and that person would be out pounding the pavement. But nobody bothers to verify computer literacy during hiring. And after that, nobody dares to tell the person who got their job through fraudulent means that they had better shape up or be fired. Instead they dump it all on the IS department.
What’s worse it the kind that act totally helpless when it comes to running the applications needed to do their work. They whine that they can’t figure it out.
But they had no trouble at all downloading and installing the bloatware screensaver, the spyware, the gambling programs and the porn on company time.
I’m old enough to remember a time when some people (aside from New Yorkers and spoiled movie stars) didn’t know how to drive. Many of their complaints would sound very familiar today in the context of computers. Was there anything wrong with automobiles’ “user interfaces”? No! It was ans still is a matter of people who either choose to do without, or to manipulate others to be their ad hoc servants. “Go to the store because Aunt Tilly can’t drive” has been replaced by “teach me to use the e-mail (for the 100th time)”.
I’m not saying that user interfaces have room for improvement — they do. But folks, please don’t throw out a bunch of sophistry about how there must be a next generation UI real soon, without any suggestion as to how or why. Many OSS coders start writing on their own time to make something that they want. If you want something, then go do it! If you don’t know how, then show me the money. But please don’t wheedle and manipulate me just because I can and you don’t want to.
Interestingly, the idea proposed of putting a query-based interface on top of the filesystem is something which I managed to convince to myself before reading the article. I have a feeling quite a number of people have independently arrived at this conclusion after using the internet (i.e. google) for a long time.
And of course the document-oriented interface is something I learnt from reading jakob nielsen. I had to write a report on user interfaces just a while ago for some class of mine
but anyway, would there be people out there interested to join up with me and try to implement such an interface? guess the first step is to sit down and think through the whole user interface experience before writing even an ounce of code. maybe we can use GTK/Glib blah blah. I’m not much of a coder but this is something really interesting that I would be willing to put some time to.
i use macosx and rarely use the dock. instead i use an application called launchbar which implemnents discussed by the article. it’s brilliant – it’s totally a different way of interacting with the system.
you access your files or application or bookmarks just by typing their name and they open e.g. if i want to use OmniOutliner, i begin to type outl and the application appears at the top of the list.
if you have macosx i recommend you try it. the potential for this application to do other things is enormous. apple should buy it!!
(like to add that i have no relation to the company that designed launchbar – i am just one happy user).
I am interested in the idea. There would be requirements for invention on a few levels (FS and GUI services mostly) but I would love to sit down and hash it out with you and everyone else. Maybe some people here would be nice to help as well as long as it doesn’t devolve into “I want a BeOS GUI/FS interface” or “I want AMIGA OS GUI interface”. Unfortunately I can only program this in Linux (or possibly FreeBSD if I can manage to learn it) as I don’t privately own a Windows based developer system and have no intention of getting one.
Now we would have to begin with general use cases and system possibilities. We would also need to talk us into how to represent a system as graphical information in the form of document. I like this idea though. It reminds me of the revolution caused by the representing of the system as files in UNIX.
Contact me and we can talk to see if this is a possibility or just a pipe dream. And the people here who think they're all that with great GUI idea, should follow up on this. Let’s look at this. BTW I am a coder (Java/C/C++/SQL) with a strong background in userfriendly design, so maybe we can help each other a little.
>How did you all succeed in reading an article with such background and foreground. The first page gave me a headache.
I didn’t
im not sure if anyone else has bough this up but isnt that what to an extent beOS could do? you could set an attribute to a file which could be any key work with a title, so you could label important email important at a file level or all you mp3’s could have genre info etc so with the right tools you could arange playlists be saying “select mp3’s by genre punk”,
i have never had first hand experiance of this but with a few key tools and if people took the time to add attributes to all files you could have a file manager, where in a filter field you could type word/words and only the files with those key words are shown?
I agree that most of the article was based on loose opinion – it would be very odd if all my apps had exactly the same UI! But there *is* a lot to be said for being able to change the UI on an individual basis. If you are using Windows2000/XP then you have some very interesting options for changing the way your desktop works. Check out the stuff that Stardock are doing with DesktopX and the upcoming NG [Next Generation?] product…
http://www.stardock.net/media/dx_mediaguide.htm
Having used GEM, OS/2, Windows, System 7, BeOS etc I’ve managed to get 99% accurate representation of these systems on my PC. Enough to be able to use all my *standards* based applications in the best way for me. I think the future is going to be based on an OS that enables each user to ‘choose’ the UI layer whilst keeping the OS and applications they need to work with. Would I pay Apple for a true OSX ‘theme’ for DesktopX on my PC? Yes. Commercially feasible for me, although probably not for Apple!
If someone likes to comment on interface design, they should take care of the website they post on !!! The colors are horrible for the eyes, not to mention the font which is suited for printed text but not for screen use.
BTW this is about the Jarred’s site not OSAlert which is ok.
“Not easy. My hunch is that a very radical redesign is required. More so than changing the arrangement of icons, menus and
dialogs. More so than the rather naf, microsoft 3d gallery approach
posted earlier… Anybody see anything truly redical?
”
An interface for doing what? Much of the discussion about interfaces
assumes that all people do on computers is file management and text
editing.
An interface which is ideal for file management and locating simple
data is totally unsuitable for constructing a 3D animation for a film,
or a complex piece of music (maybe synchronised to the animation), or
designing the electrical wiring for an office block.
Also, two programs may cover a similar area (such as music
composition) but from a different point of view. Often, getting to
know a program is getting to know the author’s philosophy. This must
be reflected in the program structure and interface.
Keyword filtering? Are you going to make me enter a key word for every picture, for every mp3 I have, for every damn file? That’s not intuitive, nor is it ‘next generation interaction’ – it’s just plain hard work!
The only part I agreed on were the comments on Evolution and Outlook, but as I understand it Evolution isn’t an Outlook clone, it just looks similar? Perhaps that’s because I haven’t used it. Perhaps I should assume the author has done his research. I’m just feeling half empty at the moment.
What was slightly annoying about the article was that there wasn’t much productive in it other than the keywords talk. There were no great suggestions on how UIs should be better.
PS. Isn’t in amusing how the first post in these comments was by somebody who obviously didn’t even read the summary properly, let alone even look at the article. Somebody once told me you should always count to 10 before you say what you’re thinking, to avoid saying dumb or offensive things. 1… 2… 3… 4… what was I thinking again?
Mike: “why doesn’t some bored college programmer sit down and write somethin like this for linux (easier said then done). i mean how feasible is it for that to happen?”
And so they did…
http://www.3dwm.org/
Am I the only person who found kai’s power tools completely, utterly and flabbergastingly unuseable? I would fire the program/plugin/etc up and spend 30 minutes trying to figure out the “intuitive” UI, all the while wishing they’d just use native system controls and respected the system’s notions of use & process. usually after 30 minutes to an hour of trying to figure out “do I initiate X by draging Y to Z or do I left-click on P or do I …” (because of course the UI’s so intuitive instructions would be superfluous) I’d just close the program and never think about it again.
I’m tired of hearing these self-proclaimed UI gurus yelling about how if all OSes were to look and feel like something Kai might have written, we’d all be swimming in some zen-like state of computer nirvana.
The concensus from these folks is that WIMP is outdated and unusable. Well, yes, WIMP is old, but so is the wheel. Just like futurists used to say we’d all be in hovercraft by the year 2000, the UI gurus always talk about the next UI step; but to the credit of the old 50’s futurists, at *least* they suggested hovercraft as a replacement for wheels. So far, no UI guru has actually had the ability to suggest something that works better than WIMP on a 2D display. At best they describe vagaries of “file stacks”, “rooms” and better search mechanisms. Woo-effing-hoo. You could do that, but they haven’t actually described a non-WIMP. They don’t describe how you’d actually *use* it differently.
My feeling is that certain mechanisms, like circular wheels & WIMP guis, work for their context and medium. Wheels are great for flat surfaces, as is WIMP, when the flat surface is a screen. When we have true holographic projectors or socket-in-the-back-of-your-head computer interfaces, I’m 100% certain some new kinds of interfaces will emerge, with the most general, flexible, and conservative ultimately winning.
I know not everyone liked Be, and some will say, just admit that it is dead, but BeOS had the filesystem queries that work similar to how Jared would like.
Another reason why interfaces are bad from a corprate perspective is ppl claming to be UI experts that end up worrying more about the eye candy than actual usability.
Example our intranet just changed and a easy to find feature was the company phone listings. With the new intranet it took me 10 minutes to find the location of the listings. The old interface I found it in 10 seconds. Another problem is that the majority of the help/content is placed in flash that has way to much eye candy and a 30 second wait to actaully get to the content, f you can find it.
Just my 2 cents
Another reason why interfaces are bad from a corprate perspective is ppl claming to be UI experts that end up worrying more about the eye candy than actual usability.
How true! I’m sick of the eye candy stuff (that looks stupid to me anyway) that I can’t figure out. What Goth widget does what? Who knows. Who cares! I’m not using that swill…
——
As an aside, what’s all this talk about search UIs? Don’t y’all realize that most modern desktop UIs already include a search engine? If you want to use that primarily, go right ahead! I predict that after awhile it’ll become obvious why organizing your data in the first place is the smarter way to go.
“If it ain’t broke, don’t fix it.”
Yes, I suppose then we would be typing by candelight, using an outhouse and traveling to work by horse and buggy.
This product features a lot of what an enhanced GUI should look like, IMO. It is a high-quality commercial project that has lots of real-world usage and feedback in its design. I find it to be very practical and intuitive. It is basically a general knowledge-management tool and an alternative/replacement to the Windows file-explorer.
http://www.thebrain.com/
No, I don’t own any BrainCorp stock. I am just very impressed with the way this tool works. Do yourself a favor and check it out.
I just finished reading Jef Raskin’s book The Humane Interface yesterday. He is a really bright guy. I think his ideas might be right on if someone were to go to the effort of implementing them. But it would be a lot of work and more importantly, a lot of planning (which is usually much less fun than programming). It is trivial to build apps as they currently are because of available toolkits, but if one wanted to redo the whole idea from scratch, there would be massive amounts of work redefining what the basic elements of the interface are.
I know how to use a computer, so I rarely notice when I am inconvenienced; however, helping friends and family always makes it clear that the current state of affairs has a lot of room to grow.
I use OS X and love the stability, graphics, etc. but the classic Mac OS had one fundamentally brilliant concept that everyone could grasp — the idea of the hard drive as a file cabinet (you keep virtual folders in your hard drive just as you keep real folders in a file cabinet) and your desktop as, well, your desktop (you can have folders/documents scattered around on your virtual desktop just as you can have folders/documents scattered around on your real desk). And, just as brilliant, you can move things from your hard drive to your desktop and vice versa just as you move things from your file cabinet to your desk and vice versa. For me, in 1987, this was a stunning breakthrough, and the main reason I’ve been an Apple loyalist ever since. Also, whenever I used this metaphor to show someone how to use a Mac, their reaction was “Now I understand.”
for one.. i think the best form to represent files as. VISUAL objects. and it is already implemented that way.. and i never see them moving away from it. sure.. they could make voice command.. but saying move this file to the desktop takes alot longer than click drag release.
sure the way its organized could be a little more intuitive. perhaps microsoft should change the name of their my documents folder to home. or something like that.. cause honestly, storing images, music and other stuff in my documents doesnt make sense to me. yes i know you can change the folders name to home.. (i did)
“Yes, I suppose then we would be typing by candelight, using an outhouse and traveling to work by horse and buggy.”
MattK, you miss the point. Each of the old technologies that you mentioned had obvious drawbacks. All have been improved upon. But they’re all pretty much the same technologies over time. We still type, light, crap and travel by wheeled carriage. Thus my point.
If we were all using voice input, bioluminescence, wasteless foods and flying cars, then your point would be well made.
I spend most of my time behind the computer and I can tell you something: I don’t wanna drag files, install programs, organize folders, write commands in “languages” and shit like that any more. I want to create and then go home.
As far as I see it, it partially is a matter of user interface, but other things are involved. Discrepancy between software and data must vanish; if that means that knowledge must be free then be it. Gadgets should know how to take care of themselves or at least be able to call software agent to help them. If that means that communication channels must be standardized, so be it.
Computer (other gadgets included) interface could be graphic, but the underlying system should not bet on that, neither that I’m willing to type, doubleclick or look at the monitor at all.
Be well,
Karel Miklav
> but anyway, would there be people out there interested to
> join up with me and try to implement such an interface?
Sure – would be an interesting project.
> guess the first step is to sit down and think through the
> whole user interface experience before writing even an
> ounce of code.
Correct.
> maybe we can use GTK/Glib blah blah. I’m not much of a
> coder but this is something really interesting that I would
> be willing to put some time to.
I think this is a prime example of what is wrong with a lot of the approach to creating something like this. Software systems are probably best off in the Bazaar, but when it comes to UI, at least the Cathedral clerics do some usability testing. If every user of a computer could contribute to the bazaar, there would be no problems, but an enourmous percentage of users are excluded from the programming commons (for good reason – you don’t need to know how to fix up a transistor to use a TV).
Start with a defined problem – something like “users want to create a 3D animation for a film” – and procede from there, not even thinking about possible technology candidates until you’ve worked out what your users actually need. Otherwise you’ll be stuck in the philosophy of your chosen technology, not choosing a technology to fit *your* philosophy.