Slashdot Mirror
Edward Tufte Weighs In on Apple's iPhone
An anonymous reader writes "Via Daring Fireball, a post from design guru Edward Tufte's site discusses his views on the interface used by the Apple iPhone. The post includes a video presentation by Tufte on the subject of video resolution on the phone. His argument is primarily that while the iPhone does a lot of things very well, Apple hasn't quite realized the platform's full potential by making screen real estate all it could be. "
Four minutes in, it's already Slashdotted. Too bad because I am going to get an iPhone in about three hours.
Sure, it's easy to say, with 20/20 hindsight, would could be better or different, but unless he's privy to all the design trade-offs which were invariably made, then I'd say the product is probably as good as it could have been, given the various pressures. Besides, it's always easier to critique someone else's work than create something novel yourself.
Chinese saying -- step too far, fall on face. A little more familiar is the phrase "perfect is the enemy of good". Attempting to release a 1.0 product that has everything absolutely perfect and without compromise is the surest way to never ship.
Perhaps iPhone2 will address some of these issues?
I think Tufte would agree with you. His opening remarks at least (I'm still waiting for the video and illustrations to make it here against the slashdot stream) seem quite favorable to the device.
Kudos to Edward Tufte for providing good detailed constructive criticism on the iPhone, including specific examples of improvements. I hope that Apple will pay attention to the FREE ADVICE that Tufte is giving and incorporate it into their next iPhone firmware update. I am sure that the UI advice that Apple pays for is likely not as good.
Personally as a product developer myself, I would welcome such good detailed constructive criticism for free from a UI guru such as Tufte. Remember that there are all innovation is based on prior innovation, so it is good to have analysis done on existing products in order to improve on future versions.
BTW, on a side note, I hope that someone at Slashdot deletes the offences racist postings above.
...where the name of the game is cramming as much information as possible into a small amount of space. Paper's dominant limiting factor was space. He says the iPhone's stocks widget could fit more information on-screen than it does. He criticizes the web browser for not using transparent buttons that would let the user see information on the web page through them.
But with dynamic displays, the game is all about minimizing the amount of retrieval time, not space. Users can tell the computer to pull up a graph for a new stock, or scroll the page downward with their finger to view the info under the buttons, or completely off-screen, with minimal effort. The biggest limiting factor is interaction. Let's keep the buttons visible, because they enable far more information than they hide.
If we sacrificed usability for screen real-estate, we'd end up with marginless documents and 4-pixel icons, which incidentally would look like windows mobile.
I enjoy Tufte's I.D. lectures quite a bit, I went to one last year and it was very informative.
I liked the video as well, with a couple of exceptions:
1) In the video, Tufte has to bust out his Sparklines (the infographics that look like lightning bolts that he mentions in the section on stocks.) He claims these have thousands of pieces of information in them but the reality is that they're merely zig-zags. As the inventor of the sparkline, Tufte thinks they're the be-all and end-all of I.D.
2) I found it hilarious when Tufte showed how he would redesign the Weather program to show more information. He said something on the order of, the only bad information design is that which leaves out important information. Sorry, holmes, I don't need to see a time lapse of cloud patterns. The Apple weather design is elegant and succinct, yours is crowded, ugly and excessive.
If you look at his examples, his primary argument is that you can cram more information on the screen because of the iPhone's high resolution. I can't agree with him that this is a good idea.
Part of the reason that people BUY the iPhone is that it's simple and stylish, rather than the existing information heavy devices like Pocket PC phones. In particular, look at his example about the Weather- Apple's widget is small and sleek. It shows you the vital information, and it does it in strong fonts and bold styling. It's clear, and it's easy.
He squishes all of that information into a tiny corner, so that he can add a large repeating satellite view- Sure it's useful in some cases, and it's certainly a neat demonstration of the iPhone's abilities, but it fails when it comes to the task of quickly giving me the important information.
It makes me squint to see the tiny version of the temperature, and shows off, rather than helping.
Sometimes developers fall into the problem of working so often because they can, not wondering if they should.
Note- He dismisses this argument, saying that information density isn't the problem, it's laying it out clearly. I agree with him in general,in that complex information can often be presented simply, but in most of his cases, increasing the density would diminish it's usefulness.
Colin Davis
The human eye can resolve 1200 dpi from 10 inches.. the iPhone is at merely 160 .. .. why do decent laser printers start at 1200 dpi?
... any more further and you probably cant resolve. (how small of an area does 1024 pixels take on your high resolution monitor .. and how high do you have to hold you iPhone away from you to have it cover that distance).
.. especially if it means I can see whole web pages without side scrolling. So I am hoping that the next iPhone has a pixel width of 800 (ie, double) or greater .. hopefully 1024.
You dont notice the iPhone having low resolution if you hold it at about 2 or 3 feet away. But in theory you can have a usable iPhone with the same size display and get 1024 pixels of width. Probably, you would have to hold the iPhone about between one and two feet away from your eyes to appreciate it though
I don't mind holding my iPhone less than 2 feet from my eyes when I web browse
How is it exactly that, in the same page where he tells us "Better to have users looking over material adjacent in space rather than stacked in time.", he puts most of his information in a fscking video?
Yah, and he's not the only one who's come up with a neat idea that isn't really as widely applicable as he thinks. He's also really not understanding the capabilities of interactive interfaces... rather than throwing all the information on one page, you drill down from the summary into detail.
For example on the stock market page, drag stocks over each other to compare them, dragging a stock all the way to the top of the page would give you more information on that stock and let you drag the screen left or right to get other stocks, flip it sideways to get the graphs, and drag left and right to compare with other graphs.
On the weather page, use the same approach, and flip sideways to get the weather map, drag up and down to see different maps.
A video screen isn't a static device, and you don't need to cram data into a single static view. Data clutter is as bad as administrative clutter.
http://www.edwardtufte.com.nyud.net/bboard/q-and-a-fetch-msg?msg_id=00036T&topic_id=1
Off topic, yes. Troll, no. That is how a Democracy works, and how our (slashdot's) semi-Republic works as well. Obviously there are opposing views here, and they get sorted out by how many individuals (who happen to have power at the time) vote or veto each idea. It is entirely possible to have a +5 Informative or +5 Insightful racist post, if it was informative or insightful to the other racists.
I've learned my lesson, though, anything attached to a bad post is automatically seen in a bad light, no matter the content. I'll just use the AC to inform the trolls or troll-questioners.
Tufte makes a good point about the hidden potential of iPhone's brilliant display. But I feel the answer lies less in resolution, and more in depth. We have been exposed to much web content that is layered (for example, pop-up windows that appear on top of existing screens that fade into the dark) that we can now discern depth on a 2D picture provided it is clear, sharp and bright. There is this 3D real estate that is not exploited in iPhone (and something that it is quite capable of).
As an example, I sometimes find it a tad annoying to keep going to the Home screen on the iPhone when switching between applications (typically when I am viewing a website and quickly need to look at maps). A dock with all Home icons down the side that appears overlaid (and magnifies each icon on fingerscroll, just like on a mac) would eliminate the intermediate step of going to the Home screen. To take it a bit further, the Maps can open in a 75% window on top of the Safari, so we can get back to Safari by one fingerstroke (Tufte's idea would be to open two windows each 50%, because there's resolution). This is, as you can see, nothing new - just something that iPhone doesn't currently have but can quite possibly do.
My sig has been answered.
Grab the coral cache mirror of the page here: http://www.edwardtufte.com.nyud.net/bboard/q-and-a-fetch-msg?msg_id=00036T&topic_id=1
Also, I've mirrored the video, as that was the slowest loading element of the page, here:
http://g.appleguru.org/iPhone_Resolution-desktop.m4v (58MB)
appleguru.org
Anyone else notice he's put "WiMax" as the operator identifier on his iPhone / iPod Touch? Seems a bit weird.
I don't have an iPhone, but if you look at the video provided, his network provider says 'ET 3G'. I thought that there was no 3G iPhone yet. Is Mr. Tufte privvy to pre-release products, or is that just a generic network identification that doesn't reflect how the data moves?
So you came to watch?
Your assumption that the time taken to select, load, and display new information is minimal not only is false, but laughably so in the case of anything operating over a cellular network.
When you are trying to browse a web page on a screen that is an order of magnitude smaller than what the author expected, it is absurd for a full 10% of that precious space to be permanently devoted to a mere 4 buttons, only one of which sees frequent use. In the case of the stocks, once the user has selected what they want to know about (be it a single stock or a set of stocks) it makes sense to display as much information as possible about them. After all, the user has already asked for the information. The only reason to leave relevant information out is if it won't fit without sacrificing the readability of the report. Tufte has never failed to understand that point, and he certainly didn't leave it out of TFA.
You are right that paper's primary limitation is space, and that this is not the case with digital displays. This is not because the digital displays are less limited in space (they never are, and in this case, the computer display is downright tiny). The reason is that the resolution of digital displays is so much lower than that of paper that the overall size doesn't really matter anymore.
The whole point of the IPhone is to be dead simple with out clutter.
now people want to clutter it up.
Im a gamer, not a grammer major. This post is full of spelling and grammer mistakes.
I am only 30 but I had trouble reading the text in the weather page he mocked-up. Maybe it was the compression in the video, but I much preferred the larger text.
I've been a Tufte fan for almost 20 years. I was introduced to his work, The Visual Display of Quantitative Information, while an Architecture student and multimedia developer in the early 1990s. This man is to graphic design what Newton was to gravity. He really defined the rules and explained why and how they are applied, technically. His statement, "Thus the iPhone got it mostly right," is him basically saying, "It's great, but I would have done some things a little differently." From a man that knows what he knows, that's the highest praise any contemporary could ever hope to get! I don't mean that sarcastically, either. I'd be ecstatic if he said anything like that about my work. Of course, I'd think it was a prank, but I don't think I'm that good anyway.
I think the icon for the weather app on teh main screen should show the actual current temperature and cloudiness/weather as an icon instead of teh current static one that shows the sun and says 73 degrees. That way I wont have to click it to find out the weather. They manage to change to current date on the "calendar" icon .. so why not dynamic icons for weather?
The video is hosted on S3, Amazon's file hosting service - I doubt any mirror you can provide is going to be much better for anyone else. The discussion around what do do wit the video from the first time Daring Fireball linked to the post, was actually almost more interesting than the iPhone video itself.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
need my rain coat on the way to the mall and on the way back 2 hours later.
That's why we have those whirly, swirly weather maps with "projections'. AKA "FutureCast"
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
I like Edward Tufte's books, but he's got some kind of plugin in his page to show that video. When you try to click-install in firefox you get 'unknown plugin'. He should be on top of that.
In this corner, wearing white and silver, rabid Mac Fan-boys who won't hear a bad word said about the greatest company in the world.
And in the other corner, wearing pale yellow with red accent - the Tufte Info-graphic Fan club who won't hear a bad word said about their patron saint.
Ladies and gentlemen, place your bets.
Are you ready to rumble?
Your comment may well apply to TFA (which I haven't read), but it doesn't take hindsight to figure out that people are going to want to use this for everything, including plenty of things Apple never thought of.
So, why did they limit it so severely that the exploit is called "jailbreak"?
Oh, and by the way, your Chinese saying doesn't apply here, either. It takes less effort to leave it open (and refuse to support custom apps, if you must) than it does to lock it down.
Don't thank God, thank a doctor!
Tuftes mantra: To clarify, add detail. is exactly, what makes most interfaces f... up. Both his new weather and stock information examples are what one will probably see on a Zune clone soon. Tasteless clutter. The apple mantra is: To clarify, hide detail Thats what I like at the interface. I did not have the iphone interface, it is almost obvious.
Sounds like they need one of those crappy TLC shows to make better use of their available space. Maybe a digital throw pillow?
"Teach a man to build a fire, and he's warm for a day. Set a man on fire and he's warm for the rest of his life."
If your iPhone is jailbroken, there is this Dock application: http://www.iphonehacks.com/2007/11/iphone-dock.html It's not perfect (a bit too responsive to scrolling, icons too small, can't change the order of the icons), but it's definitely functional, and I hope Apple implements something similar in the future.
Although to be fair, if your stock market widget is going to start pulling down 14,000 datapoints and assembling them into a graph, or your weather widget is going to start displaying complex radar images, then it's quite likely that retrieval times are going to be substantively worse. The key word that you're ignoring is "relevant". I doubt many people will find all 14,000 datapoints of relevance when looking up a stock price on their iPhone. They're probably only interested in a general sense of the trend. Kind of like asking how old someone is and getting a reply accurate to the number of days, it's not clear that more information is always more clarity, despite what Mr Tufte says -- at least for me.
paper wins!
It's true that dynamic displays are different from static ones. But there's another reason besides the staticness of paper that you might want to put a lot of data on the page/screen simultaneously: you might notice something interesting about the data if you can see more if it at the same time.
They are mini graphs, and I'd find them highly useful. I get annoyed having to go through and touch each stock to see how it has been doing. Sparklines would give me the flavor for all stocks at once.
I don't agree. Apple's weather app does not provide a real weather report. A pretty snow icon is nice, but is that 2 inches or 2 feet?--that just might affect my plans for the day. Does the rain icon mean showers or torrential downpours? Is it calm or gale-force winds? Tufte's design provides an actual weather forecast. The radar image is less useful, but still nice. Especially for those who don't life in the center of big cities, it can give a better idea of what to expect where you are. And a glance at the radar often gives you information about how severe the weather is, how widespread, and how fast it is moving that can be hard to glean from a text report. But the image should load last, because often I don't need it, in which case I don't want to wait for it to load just to read today's forecast.
Hopefully, if Apple doesn't step up a 3rd party developer will, once Apple releases the iPhone SDK.
I'm surprised and disappointed to read comments from people who've read Tufte's books and agreed with them, and then say the opposite about this case. I wonder, did they ever 'get it' ?
Tufte's view, consistent for decades, is that the information display should be designed around the human visual system's abilities and preferences, not the designer's prejudices or what's easy for the display system.
The human eye automatically "drills down" in an information-rich visual field by focusing the fovea on anything that is noticed as being of interest. Further information on the subject of interest is gained in a dozen milliseconds by the act of focus. No jumping to new pages over a second later.
A couple of posters offered the absurd assertions that
a) Tufte is stuck in the paper era - when he's been commenting on computer displays for 20 years. His criticisms of the screen real-estate forgone to 'computer administrative debris' in Mac and, later, Windows, go back to their inception.
b) That space is limited on those paper pages when they are far more information-rich than screens. Multiply 8.5x11x300x300 and get over 8 Mpixels, guys. (And an open magazine is twice as big; an open newspaper, 10x that!) Why do you think most people prefer to read on paper even now? Richer colours, too; compare TIME print edition photos to the web pages printed out.
People who think information-thin combined with drill-down is the way to go are responsible for those frustrating answering-machine menus.
And definitely have never taken a look at Craigslist, where there are a maximum of index words per page, using smaller print, and every piece of information in the index is also a 'control'. a link to another information-dense page. You rarely have to go more than three clicks in until you are looking at a list of the things you want, out of all the country and all the products and services there are.
Bottom-line: provide the user with as much information as possible, use visual cues (size, colour, position) to prioritize, and have trust that they will pick out what they want. Providing them with less information so as to lead them by the nose down your little trail insults their intelligence and human abilities.
I get the distinct feeling that Tufte understands data visualisation, but not interface design. These are different things, and he's letting his expertise in one area make him think he can make pronouncements from on high in other areas and comes out with some real bullshit as a result.
His "to clarify, add detail" rule could be applied to his comment on the photo browser. He says they should be grey not white, and only one pixel wide, but gives no reason why. I'd like some detail to clarify why he says that! It would not fit more images onto the screen, it would add no information content, it's barely even an aesthetic change to the design. It's news to me that arranging images against a plain white background is a bad approach. I've met a lot of smart people that like to "show off" by making detailed comments like this, without any actual substance or empirical evidence to back up what is simply their own preference. Tufte seems to be doing so here.
He criticises the stock app for being "cartoony" and "PowerPoint" like, which seems again a mere preference rather than an objective comment, uses words designed to provoke an emotional reaction rather than an intellectual one. He claims his app has more detail - which of course it should when it only has three stocks, not six. But I don't see how x thousand points of data points in a tiny little graph is of use. First of all, if you fit thousands of data points into a single graph, it's going to need a damn big piece of paper before I'm capable of distinguishing them, combined with a ruler and a set square if I want to get the value for a specific data point. Second, why would I want this level of detail on a phone app? Personally, I find the iPhone's red light / green light view combined with percentage points useful - it jumps out at you when e.g. the market crashes as it did recently. In Tufte's example, it's impossible to tell what recent market changes have taken place, and there is no obvious way to quickly see data for e.g. the last week. The "modest data graphic cartoon" conveys just as much information to the viewer as his "image resolution" with thousands of data points, and is the kind of thing a portable stocker checker would be used for. Tufte is letting his expertise get in the way of understanding the use case - all his catchphrases are there for the converted, but his use of them here just annoys me.
Here's a nice little piece - take a look at his site at http://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=00036T. He criticises the iPhone browswer for having 10% of the screen used for buttons, but in his own designs he comments "about 90% of the image is substance". Clearly he's happy with that 10% sacrifice when it's his own work. And if you look at the designs, you'll note that in each case there is a navigation bar of some form at the top or bottom of the page. What a hypocrite.
Finally, he's very keen on getting rid of computer admin debris. The problem is, he treats looking at a web page the same as looking at a picture. But when I'm looking at a picture, I don't want to bookmark it (it's already in my collection), and I don't want to make a webclip of it. I don't need the back button with photos, because I can navigate via the photo collection. But I do need those functions in the browser, and I need them large enough to easily hit with my finger. We're all used to scrolling down webpages, so having a mere 90% of the screen available, and an intuitive flick of a finger to scroll down, is perfect. Commenting that the button bar should at least be transparent strikes me as just one of those condescending little compromises some people like to make when they know they won't convince the other side of "the right answer". It would be bad interface design to have application buttons hovering over hyperlinks, making it distinctly ambiguous what would happen when you touched that bottom 10% of the screen.
In particula
Edward Tufte has done a great deal of novel and ground-breaking work, and has done a great deal to share his insights w/ others in the field, starting w/ his seminal _The Visual Display of Quantitative Information_.
For my part, ``good enough, isn't'', and I far prefer the Zen parable of the archers --- three archers compete for a prize, all strike the mark, a fish, and are then asked ``At what were you aiming?''
The first answers, ``The fish.'' as does the second, but the third?
``The center of the fish's eye.''
You can't be any better than you try to be and I'd much rather wait for the efforts of a person striving for perfection than accept those of someone willing to be mediocre.
William
Sphinx of black quartz, judge my vow.
No interaction is almost always preferrable to interaction. If you can display the information the user wants, display that information instead of making the user act to see the information. Every user interaction is a source of errors.
That is generally a bad idea. Interaction should be considered as a compromise if not other solution can be found, not as a good idea in itself. Every time users have to interact with your application, a certain amount of users will fail. If you can show every information the user has to know without him interacting with your application, you've killed that source of errors.
You should read this paper on why interaction is bad, and how interaction can be avoided.
Quote:
The ubiquity of frustrating, unhelpful software interfaces has motivated decades of research into "Human-Computer Interaction." In this paper, I suggest that the long-standing focus on "interaction" may be misguided. For a majority subset of software, called "information software," I argue that interactivity is actually a curse for users and a crutch for designers, and users' goals can be better satisfied through other means.The only 3 things I look for on the web re weather are
1) 4 day forecast in the morning
2) radar if its raining to decide if to go now or later (people with cars dont care)
3) hourly temps for the past 24hrs
If in doubt, make a damn advance config option to specify "are you a technical geek and want more techy info" how hard is it for dynamic screens, static designs are so 1985 like DOS.
Liberty freedom are no1, not dicks in suits.
I agree with Tufte that the buttons at the bottom of the screen in Safari take up too much real estate, and I like his solution, having them transparent. They could fade after a minute or two, with a tap at the bottom of the screen bringing them back to full intensity.
I also agree that the iPhone can't afford to leave the address bar at the top of the screen like a PC browser, and that Apple made a good choice in having it scroll off with the page. But it does highlight a problem: With very long pages (many blogs for example), it can take a very long time to scroll to the bottom of the page, and a long time to scroll back up to the address bar. Of course, you can use the forward and back arrows at the bottom, or the pages button to call up a new browser page, but that's kind of awkward.
I'm hoping Apple will bring the 3-finger swipe of the AirBook to the iPhone. 3-Finger swipe down to jump to the bottom of a page, 3-Finger swipe up to jump to the top and the address bar. Right and left swipes could do forward and back, and make the fading of the bottom buttons less of an issue.
This is true but easily dealt with by prioritizing retrieval. First display the most important info (current stock prices, today's forecast) then fill in the additional information that some people will be wanting.
Part of the reason I bought it (and one that has been promoted in Apple's ads) is that the browser does not display a "simplified" version of the Internet as do most cell phones--it shows standard web pages with all of their detail and complexity.
It is understandable that Apple was in a hurry, and simply implemented Dashboard widgets on the iPhone, but an iPhone is not the Dashboard, and iPhone apps can handle more information.
-Ted
-=-=- Quantum physics - the dreams stuff are made of.
There's a classic Dilbert cartoon that shows some sales guy trying to sell Dilbert a non-interactive computer. It's easy to use. "There's only one button, and we push it at the factory." "What does it do?" "Whoa! Let me get an expert!"
If you just want to make phone calls, the iPhone itself is a waste of time. I want a phone that has only a keypad and a small monochrome display, with no games, no internet access, no maps, no weather, no music, no ringtones, just a damned phone. I have a separate PDA that isn't tied in to a phone company, upon which I can install any software I want, for that kind of thing. I can leave my PDA behind when I'm going someplace that's hazardous to digital devices, and still carry my phone. I can leave my phone behind without losing my personal information.
So I will agree, if you don't need a lot of capability, you don't need an interactive device.
But the iPhone isn't for someone who doesn't want a lot of capability. It's for someone who wants more than you can fit on an index card. You have three choices: limit the amount of information, add clutter, or let the user ask for more information. The first and third choices are both valid. What Tufte's arguing for is the second.
Clutter can be fixed not by throwing away information but by changing the design.
But his video doesn't show that. It shows more clutter.
Look at his proposed weather interface. It's more cluttered than the original.
Look at his proposed stock interface. It's not only cluttered it's actually got LESS useful information than the original. It really doesn't have the thousands of numbers he thinks it does.
I suspect you misunderstood the article you directed me to.
The widget he built is highly interactive. It only shows arrivals, departures, and trip times, and hides details like the map and other schedules until you request them.
The article is about selecting the information that you provide in the overview so that you don't have to drill down most of the time. Not about cluttering the overview with extraneous details the way Tufte was.
Tuft made a video! The horror!
There is no indication that the parent post read the article, or even the summary; just from the title one could compose such a post. The same post could be made on articles panning/praising any product -- just change the subject of the last sentence from "iPhone2" to "Vista 09" or "OLPC 2.0", etc.
Yay, moderators.
-- @rjamestaylor on Ello
By chopping up my remarks for your reply, you've missed answering the main points I was making:
a) the more info you want to send, the slower it will be. Retrieval times are non-trivial for wireless gadgets at present, as I know to my cost in using my blackberry, so it's sensible design not to send more information than most users need.
b) even if you present data clearly, concisely and efficiently, sometimes more data will obscure rather than clarify because it just gets in the way. When most people ask the question, "What's the weather like today?", they want a simple answer like "sunny" or "cold with showers from time to time". A six-frame animated gif of complex radar images is just extra noise obscuring the signal. I think sparklines are a clever idea and very powerful for analytical work, ie measured reflection on a complex dataset to come to a thoughtful conclusion. But you'll have a hard time convincing me that most users, most of the time, want to do analytical work on their iPhone, as opposed to look at whether the market is up or down or if they need an umbrella this afternoon.