Release 2.0 Issue 2.0.4, August 2007 http://r2.oreilly.com “While many information visualizations are amazing, fun, and often quite useful, some of them promise to improve society They may help us change our minds, collaborate, make better decisions, and re-knit a social fabric that has frayed badly over time.” Jerry Michalski, from The Implications of Visual Literacy, page Release 2.0 Issue 2.0.4, August 2007 ISSN 1935-9446 Published six times a year by O’Reilly Media, Inc., 1005 Gravenstein Highway North, Sebastopol, CA 95472 http://r2.oreilly.com This newsletter covers the world of information technology and the Internet — and the business and societal issues they raise executive editor Tim O’Reilly tim@oreilly.com editor Jimmy Guterman jimmy@oreilly.com publisher Sara Winge sara@oreilly.com art director Mark Paglietti markp@oreilly.com production design Drew Barton drew@oreilly.com copy editor Laurel Ruma contributing writers Jerry Michalski Peter Morville David Weinberger Contents 01: Information Visualization: The State of the Art If you want to understand what today’s never-ending onslaught of data is struggling to reveal, you need to learn how to look at it By Jimmy Guterman 03: The Implications of Visual Literacy An annotated guide to getting what you see 09: By Jerry Michalski What I’ve Learned from My Brain Ten years inside an information visualization program By Jerry Michalski 12: Improving the Interoperability of Online Visualizations Notes toward the next generations of mashups By Jerry Michalski 22: 16: Visualizing Forward The four paths information visualization is leading us down and where they might lead By Jerry Michalski 19: Still Tomorrow’s Technology Information visualization is great, or it will be, if it ever comes By Peter Morville 21: Caution: Low Visibility By David Weinberger Does anyone remember words? The Number: Vista Perception Critical Update Microsoft’s new operating system is a bust, the pundits say/wish Don’t believe ‘em © 2007, O’Reilly Media, Inc All rights reserved No material in this publication may be reproduced without prior written permission; however, we gladly arrange for reprints, bulk orders, or site licenses Individual subscriptions cost $495 per year 70557 subscription information Release 2.0 PO Box 17046 North Hollywood, CA 91615-9588 http://r2service.oreilly.com customer service 1.800.889.8969 1.707.827.7019 r2@oreilly.com By Jimmy Guterman 24: Calendar Jimmy Guterman is editor of Release 2.0 and editorial director of O’Reilly’s Radar group You can reach him at jimmy@oreilly.com Information Visualization: The State of the Art Do you want to understand what an avalanche of data is trying to show you? Chance are you have to find a clever way to look at it If you spend much of your time in meetings, chances are you’ve been confronted with a slide that attempts to tell a story It might look something like this: In Peter Norvig’s witty reduction of Lincoln’s thrilling Gettysburg Address to a soporific PowerPoint deck, he illuminates how the wrong image (in this case, a chart representation of “four score and seven years ago”) can eliminate all meaning and style from the presentation of information And the examples we see in corporate conference rooms now are less funny but just as useless One of the few positive side effects of this era’s information overload is that we’ve found some new and useful ways to organize the avalanche of information that drops onto us every day Roughly a quarter century after the publication of : 0 Information visualizations are not about pretty pictures Indeed, the ones that are most decorative are often the ones that yield the least useful information 0 : Edward R Tufte’s instant classic, The Visual Display of Quantitative Information, we’re still in the early stages of learning how displaying complex data in clever, clarifying ways can increase understanding and improve resulting decisions Information visualizations are not about pretty pictures Indeed, the ones that are most decorative are often the ones that yield the least useful information to share and act on Much of making smart business decisions is deciding which data sources are trustworthy and which visualizations of those data sources tell stories that clarify and reveal what’s behind the always-moving, Matrix-like walls of numbers Achieving visual literacy helps business people decide which stories and believable and make better decisions In this issue of Release 2.0, our all-star lineup examines the state of information visualization, how it got here, and where it might be going Jerry Michalski returns to this newsletter after a much-too-long absence He points to some of the most probing examples of information visualization nowadays, details his own decade-long involvement with one InfoVis program, senses some early signals on how these disparate visualizations might one day interoperate, and delivers his own wish list for the future (It’s an article meant to be clicked on, so you may prefer to read the PDF version that comes with your subscription) Peter Morville, author of Ambient Findability, also published by O’Reilly, warns us not to get carried away with today’s InfoVis tools And, finally, David Weinberger, author of Everything Is Miscellaneous, reminds us that sometimes (think: Gettysburg Address) words may be all you need to make your point And then we look at the numbers to see if the convention wisdom about Microsoft Vista adoption is based in external reality Are you using information visualization tools? Have you found ways to dig out from the information avalanche? Please join us on the O’Reilly Radar blog, at http://radar.oreilly.com, as we continue the conversation nn Release 2.0.4 August 2007 The Implications of Visual Literacy Jerry Michalski The Implications of Visual Literacy They’re everywhere Where are the best ones and what they do? by Jerry Michalski If you’re a fan of visualization, these are fun times Elegant, useful visualizations abound online Mash Google Maps together with Craigslist’s housing listings and you get HousingMaps (http://www.housingmaps.com) Add Twitter tweets (short messages) to some geocoding information and a map, and you get Twittervision (http://twittervision.com/), which animates tweets popping up around the globe as they happen Want to see how stories plop into Digg, as they plop? Try Stack from Digg Labs and Stamen Design (http://labs.digg.com/stack/) Which candidates are where on the campaign trail and what did they just say? Check Map the Candidates (http:// www.mapthecandidates.com/) Want to track where all the man-made satellites are now in orbit? Go to NASA’s 3D tracker (http://liftoff.msfc.nasa.gov/RealTime/JTrack/ 3D/JTrack3D.html) and give the display a little tug, to rotate it Jerry Michalski helps companies develop strategies that build relationships with their customers as well as among their employees A longtime technology industry analyst, he is best-known in these parts as an editor for Esther Dyson’s Release 1.0 He can be reached via http://sociate.com/ Twittervison superimposes Twitter messages on a map in either 2D or 3D : 0 Release 2.0.4 August 2007 The Baby Name Wizard NameVoyager will let you know just how mundane the name you picked really is, or isn’t 0 : The Implications of Visual Literacy Jerry Michalski University of Maryland’s Ben Shneiderman invented “treemaps” years ago, but Martin Wattenberg’s Map of the Market (http://www.smartmoney.com/ marketmap/) made them popular as “heatmaps.” Wattenberg also scored some points for visualization with the Baby Name Wizard NameVoyager (http://baby namewizard.com/namevoyager/lnv0105.html), which accompanies and enriches his wife Laura’s book Wattenberg is now at IBM, part of the amazing Many Eyes project (http://services.alphaworks.ibm.com/manyeyes/app), where participants upload datasets and collaborate on their analyses Social network services such as Facebook and MySpace are hot, as are social network mapping applications such as Fidg’t (http://www.fidgt.com/visualize), which helps you see what your friends’ interests are; PieSpy (http://www.jibble org/piespy/), which infers who is talking with whom from listening in on Internet Relay Chat conversations; and Valdis Krebs’ InFlow (http://www.orgnet.com/inflow3 html), which performs similar analyses on email data and Amazon purchase data You could go to Facebook and “friend” Episcopalian Bishop Marc Andrus, who just added a social-network visualization app called Friend Wheel to his Facebook page (so I did, too, of course) Is this too much information from a spiritual leader? You decide Then there are social analysis tools, where it’s not the network that’s social, but the method of analysis and discussion With these tools, which include Swivel (http://www.swivel.com) and Visualization Lab’s sense.us (http://vis berkeley.edu/papers/sense.us/), participants can compare, discuss and annotate data-driven visualizations built around census data I’ll return to this social sub-genre shortly The examples I’ve cited so far are data-driven There’s another important category of visualizations that are illustrative, based on concepts or metaphors Some just make a point, some tell longer stories, weaving together multiple points Their creators use many media, from animations (remember JibJab and The Meatrix? http://jibjab.com, http://www.themeatrix.com) to videos, comics, storyboards, presentations (for the equivalent of PowerPoint on the Web, see SlideShare, http://www.slideshare.net/), screencasts, serious games, flipbooks, photo essays, and Java applets Want to talk alongside your data the way weather forecasters talk to their maps? Watch a GapCast from GapMinder (http://www gapminder.org/video/gap-cast/) Want a nice how-to video? Head to Instructables (http://instructables.com) (O’Reilly is an investor in Instructables) Among the many inspirations for these projects are Edward R Tufte, the noted expert in packing many dimensions of data into one display, and Muriel Cooper, whose research into text sizing you’ll recognize every time you see a tagcloud That’s just a taste of what’s available For an impressive compendium of visualizations, visit VisualComplexity.com (http://www.visualcomplexity.com) For a more humorous take, check out the Periodic Table of Visualization Methods (http://www.visual-literacy.org/periodic_table/periodic_table.html) from the Visual Literacy site Some visualizations are designed to get people to change their opinions Some information visualizations just make a point, tell longer stories, or weave together multiple points More and more are using multiple media : 0 Release 2.0.4 August 2007 The Implications of Visual Literacy Combine open databases with engaging information visualizations and you can change how governance works More than fun 0 : Jerry Michalski While many of these examples are amazing, fun, and often quite useful, some of them promise to improve society They may help us change our minds, collaborate, make better decisions and re-knit the social fabric that has frayed badly over time For example, some visualizations are designed to get people to change their opinions Remember Ben Cohen’s Oreo histogram about military spending during the last election cycle (see http://www.truemajorityaction.org/oreos/)? Notable among these eye-opening opinion-shifters are Ingo Gunther’s World Processor (http://worldprocessor.com), which maps statistics onto the globe (e.g., violence, water reserves, debt, per-capita income), and Bruce Mau’s Massive Change (http:// www.massivechange.com/exhibition), which uses a wide variety of techniques and media in physical installations to make palpable the size and scale of pressing global issues But those visualizations are still one-way: An expert creates a compelling map or story, then sends it into the world Perhaps more interesting are visualization tools that improve discussions among peers in order to make better decisions Tools like these help us rethink important concepts such as governance and democracy The use of visualization in governance may shift over time from a view of visualizations as end products to the process of visualization as central to finding common ground and making the decisions themselves To paraphrase Martin Eppler of Visual Literacy, our focus will move from visualizations to visualizing One sub-genre of discussion aids offers argumentation mapping These tools help groups draw out the logic of an argument, showing which points support the premise being put forth and which undermine it Examples include Let’s Focus (http://www.lets-focus.com/), Compendium (http://compendiuminstitute.org, with roots back to Jeff Conklin’s gIBIS, TruthMapping (http://truthmapping.com/), and Rationale (http://www.austhink.com/rationale/) Another sub-genre, open databases coupled with analytic tools, is changing how governance works At the very local level, communities are mapping lot and tract information against crime and health data and coming up with datadriven solutions to thorny problems that broad-brush national strategies often fail to address At a national level, watchdog groups such as the Sunlight Foundation (http://www.sunlightfoundation.com) are improving the oversight of elected officials The Foundation’s Mashup Lab created Popup Politicians (http:// sunlightlabs.com/popuppoliticians/), a useful Javascript widget that offers The Sunlight Foundation’s Visualizing Earmarks shows clearly what so many have sought to keep hidden background information on politicians in context, and Visualizing Earmarks (http://sunlightlabs.com/visualizingearmarks/), which uses the Many Eyes analytic platform to map earmarked budget funds to the politicians who created them Many Eyes, sense.us, Swivel, and similar tools hold the promise of broader, more democratic participation, the way blogging brought more people into writing, digital cameras and photo-sharing sites broadened photographers’ reach, and inexpensive video cameras plus YouTube let practically anyone generate movies Some of the most compelling socially constructed visualizations aren’t analytic, but rather cartographic People are uploading GPS and geocoded data to mapping services and turning out useful maps of race routes, graffiti waves, and favorite restaurants Google has been aggressive in adding capabilities to its Maps and Earth offerings to make such customizations easier and more powerful Microsoft and Yahoo are hot on its heels With luck, we’re entering an era of critical, social thinking : 0 Release 2.0.4 August 2007 We haven’t yet figured out how to detect and defuse filter out all the demagogues The Implications of Visual Literacy Jerry Michalski The question of multimedia literacy pops up often as underlying technologies change Once it was experimental films with multiple plot lines; today it’s multitasking Millennials who can listen to music, download movies, chat with six friends, play World of Warcraft, and still finish their homework Or claim to Yet I don’t think that the new literacy is this form of continuous-partialattention-driven ADD It has more to with the reintegration of dialogue, images, analysis and collaboration Remember your first PowerPoint? A wide variety of visualizations are in everyday use To understand the expressive variety, check out the taxonomy of types of visualization that Bob Horn offers in Visual Language Visualizations are becoming commonplace in business settings Graphic facilitators make visual records of what happens in meetings and conferences Consulting firms such as Dynamic Diagrams and XPLANE help organzations make visual sense of complex systems or product offers Oculus (http://www oculusinfo.com) sells a suite of business data visualization tools Knowing which type of visualization to use, and when, remains more of an art than a science, but we’re getting better at it, much as we learned not to mix every font available in early word-processed documents, not to use eight-point type on PowerPoint slides, and not to use tags on Web pages Now those things are received wisdom, embarrassing gaffes when witnessed anew Don’t be fooled into thinking better visualizations of rational decision-making will make everything go smoother and help the best decisions rise to the top We clearly haven’t figured out how to detect and defuse demagogues We’re early in the process of mastering this new-yet-old visual vocabulary The tools don’t interoperate (see page 12) Visual literacy is not widespread We’re also still discovering how best to discuss issues collaboratively, with an online record of what transpired Just because the Web has the memory of an elephant doesn’t mean it’s remembering what’s most useful Turning random wanderings into memorable insights is crucial Perhaps sunlight is the best disinfectant, and increasing transparency coupled with better tools and collective intelligence will improve our lot nn 0 : Release 2.0.4 August 2007 Improving the Interoperability Of Online Visualizations Jerry Michalski Improving the Interoperability of Online Visualizations What if there were a way to connect all these islands? by Jerry Michalski The goal is to define an environment within which a rich variety of visualizations can coexist, but which allows for easy connections to occur between them Not long after feasting on the hundreds of visualizations at http://VisualComplexity com, I wondered: why are all these visualizations disconnected? Why can’t data from one feed the next? Why can’t I tell a story linking one to another, or add annotations to them in turn? How might we connect them so they can work together? I can imagine various ways that interoperability might play out One simple way is the mere juxtaposition of different visualizations Show me the pie chart of family expenses in Ghana next to the histogram of social program expenditures over time, for example What if I would like to comment on both, to annotate them? The same Web 2.0 properties that make mashups easy can help here, but their limits are reached quickly Although APIs and RSS feeds are extremely useful, even the more advanced features of HTML and CSS today can’t match the visual flexibility and power of Adobe’s Flash or Sun’s Java (and Microsoft’s new Silverlight) Alas, these latter environments typically lead to closed applications Yet there is progress Several of the sites I mentioned in the lead article shine here, so I’ll use them as examples The goal is to define an environment within which a rich variety of visualizations can coexist, but which allows for easy connections to occur between them So, let’s address three levels of interoperability: application, data, and social Application interoperability How would you assemble a narrative thread that ties together a bar chart from one site, several pages from an online slide show at another, and a video from yet a third site? The baling-wire-and-twine way is to write some prose in a blog entry or Web essay, then salt the entry with links to the various visualizations It’s not elegant and it doesn’t help make the point much What if the working environment, the platform atop which we built applications and illustrations, allowed us to link things together internally the way Gliffy lets you chain together Web graphics (http://www.gliffy.com), and the end product could look as elegant as one of Scott McCloud’s sequential art pieces created for the Web (http://www.scottmccloud.com/comics/comics.html) ? Don’t hold your breath for such a system It’s not in development, as far as I know Today there isn’t that much application interoperability for visualization applications, beyond data feeds and occasional published APIs 12 : Data interoperability Rebecca Xiong’s WebFan visualizes the activity at Web-based message boards The news is better on the data front Sites such as Swivel and Many Eyes encourage visitors to upload data, blend it with other data, and plot the results Others are coming For some time, it will be easier to pour datasets into and out of these sites than to connect the sites directly to one another, but that won’t be too far behind One of the major hurdles to strong data interoperability lies not in the technical data-interchange standards, but instead in the consistency of the nomenclature used to describe the data If one dataset represents data collected at the end of each period and the other’s data is collected at the beginning, plotting the two together may introduce artificial lag effects If “customer” in one dataset means paid customers and in the next means those active in the last year, you’re not comparing precisely similar things Data hygiene is essential : 13 Release 2.0.4 August 2007 Improving the Interoperability Of Online Visualizations Jerry Michalski Much of information visualization is bringing together multiple data sets to see what you’ll find Of course, without data in the first place, hygiene is pointless For years, Carl Malamud has been pulling data that ought to be public into the public view One of his early efforts was to turn the SEC’s database of filings, available then for $1 a page, into a public resource known as EDGAR (http://en.wikipedia.org/ wiki/EDGAR) More recently, he convinced C-SPAN to release much more of the video it captures into the public sphere Malamud’s latest project is a sort of data conservancy, http://public.resource org, which encourages visitors to purchase newly available materials once, thus funding their placement into the public domain There are a few other major data collection projects, notably the Internet Archive (http://www.archive.org), OurMedia (http://www.ourmedia.org), Google Base (http://base.google.com/), and Metaweb’s Freebase (http://www.freebase com) Each of these organizations uses a different set of benefits and capabilities to attract content 14 : Social interoperability The Visualization Lab’s sense.us prototype is a nice example of social interaction around data In it, participants can add markups to the charts, link those markups to their comments, and essentially conduct informed discussions around the analytic evidence One unexpected hurdle to smooth social interoperability is cultural differences in interpreting visual cues Martin Eppler of Visual Literacy recently studied students from Cambridge University and a Chinese university, and found their perceptual styles markedly different The desire to collaborate over visual artifacts quickly crosses into a part of the old groupware space that has not been subsumed by the moniker “social media.” It’s the screensharing and collaborative annotation tools that have mostly survived in the market as paid offers, including WebEx, GoToMyPC, Breeze, and Glance And even these are mostly playing what’s on one screen for others, rather than actually linking the contents of the screens in useful ways It’s unfortunate that such features aren’t simply baked into the operating system, or part of the browser architecture While it is true that Windows and Apple’s OS X each have some of these features, cross-platform compatibility is poor In an era when operating systems are less and less important, this feature gap should dwindle in significance If advanced projects such as Squeak (http:// www.squeak.org) and Croquet (http://www.opencroquet.org) had propagated sooner and more broadly, we might all have these features now In these latebinding, distributed environments, collaborative tools are either built in already or easy enough to instantiate and make available to all present Unfortunately, we remain trapped in operating systems, office suites, and application development tools that progress at a glacial pace Nevertheless, the pace with which new, useful visualization sites and applications are emerging is promising Enough enterprising minds care about this space to fill the gaps nn If these advanced projects had appeared sooner, their features might have been baked directly into our computer operating systems : 15 Release 2.0.4 August 2007 Visualizing Forward Jerry Michalski Visualizing Forward Notes on where the art may be going by Jerry Michalski The most forward-looking InfoVis shops are choosing exploratory projects over task-specific ones What paths might information visualizations take as they become more pervasive? Let’s explore a few The first path is pragmatic HousingMaps is intuitive and cool, but it’s far from finished What if you took a home-search application all the way forward? What if the goal was to make it useful for people looking for a place to live? What else might it do? I know what I’d want: ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ On the map, hide neighborhoods I don’t like; highlight streets I love Set alerts so I get an SMS notice when a property is available on a street I love Hide properties I’ve checked out but don’t like Connect photos of properties I take and upload to Flickr with the right properties in HousingMaps Same for videos Let me annotate all the above, so I can remember notable features and flaws Manage a simple decision-support table with properties and variables Help me collect data on the variables, populate the table and narrow down the choices Invite friends (privately) to see the resulting maps/pictures/table, then discuss them with me Screen-share all the above as I talk with my friends or real estate agent Enable my mobile phone to become a neighborhood survey instrument, capturing location, images and data as I move around places I like (eg., this street’s a 7; that one’s a 10! let’s make an unsolicited offer on that place!) Turn my car’s in-dash navigation system into a command center for the house hunt Many of these features could be achieved easily, the way Paul Rademacher originally created HousingMaps from amazingly few lines of JavaScript Others would take a bit more work You may desire a different set of features No problem So this first path involves taking abstract visualizations and thinking them through Despite all the buzz around human-centered design and empathic design, we haven’t really folded in the user experience We haven’t put ourselves in the users’ shoes enough It’s time to put visualizations to serious work I used house hunting as an example to focus the idea Now extrapolate into whatever domains you feel strongly about, whether it’s job hunting, sports, investing, or travel 16 : The second path is exploratory The Trulia site has an interesting, Stamen-designed application called Hindsight (http://hindsight.trulia.com/) that animates building construction in cities over time As a marker glides along a timeline, dots appear on a city map You can see your city grow up, with bursts of development during boom times fleshing out different neighborhoods Trulia’s Hindsight lets you see how your local landscape is changing over time Eric Rodenbeck, one of Stamen’s principals, explains what kinds of projects the company seeks out It wants engaging, exploratory applications, rather than task-specific workhorse apps Hindsight is one example, and the apps that Stamen has built for Digg Labs are others (http://labs.digg.com/) These include Arc, Stack, BigSpy, and Swarm—all different ways to see stories as they’re being dugg by Digg users In Arc, stories populate a circle, in Stack they drop into a line like Tetris pieces BigSpy has them as pure text, sized by the frequency they’re dugg And Swarm is a concept map view They’re all elegant and quite captivating Rodenbeck and his colleagues want to build applications that make you think, that help you create associations you might not have thought of with more conventional tools Many of today’s tools have this exploratory feel They allow people to test theories, see things expressed in new ways, and play out ideas of how the world works : 17 Release 2.0.4 August 2007 Visualizing Forward The third path is persuasive When you couple visualization with feedback loops, the result is powerful It can change behaviors To see a great example, you need go no further than your nearest Prius Dashboard The display that shows when the brakes are regenerating, when the gasoline engine cuts in, when you’re on electric power only—and how all of that influences gas mileage—is mesmerizing In fact, drivers tend to adjust their driving to maximize mileage, as if the software were a video game Sometimes it takes a person to the persuading Hans Rosling from Sweden’s Karolinska Institute has been talking about statistics in ways that promise to engage broader audiences in discussions around important issues At GapMinder (http:// www.gapminder.org/), the site he runs (and sold recently to Google), tools help animate statistics Most engrossing of Rosling’s endeavors are his GapCasts, in which he inhabits screens full of data, illustrating trends the way your local weatherman dances with weather fronts Given how popular videos, podcasts, and screencasts are today, I see some data-driven greenscreen work in our futures The fourth path is emergent and user-powered The U.K is famous for its Ordnance Survey maps (among other things), which are accurate and detailed and a great way to get around the country The only problem is that the British Government decided not to release the survey data for public access, depriving interested parties of many stripes access to excellent geocoded data In 2004, some hobbyists decided to take matters in their own hands and reconstitute the data, using volunteers armed with GPS gear The result is OpenStreetMap.org (http://en.wikipedia.org/wiki/OpenStreetMap), a growing trove of geocoding data collected by throwing mapping parties in various cities, bringing together experienced mappers and newbies As the project grew, a delivery service offered to pool its vehicles’ GPS data, greatly enriching the project’s resources As a result of this grassroots effort, a deep resource of useful data is now available free of charge, and it is moving to other countries One of the byproducts of such a collaborative effort is that it leaves behind networks of people who have bonded over a common task That means they trust one another and know how to work together They may be important networks to know about in future emergencies, or for other projects There are certainly other paths that visualization will take One light-hearted rendition of what a piece of the territory looks like is this Japanese information architecture trendmap (http://www.informationarchitects.jp/ia-trendmap-2007v2) Recently, Manuel Lima of VisualComplexity.com issued a challenge to his peers to create a conceptual map of the visualization territory (http://www visualcomplexity.com/vc/community/challenge.cfm) It will be fun to see what that produces, because the many faces of visualization will co-evolve in unpredictable ways The optimist in me believes that it will lead to sharper analyses, better decisions, and deeper understandings nn 18 : Jerry Michalski Collaborative efforts leave behind networks of people who have bonded over a common task They trust one another and know how to work together Release 2.0.4 August 2007 Still Tomorrow’s Technology Peter Morville Still Tomorrow’s Technology Information visualization is great It’ll be even better when it gets here by Peter Morville When I consider the prospect of information visualization, I can’t help but recall an ancient sign outside the Cross Gates Pub in Lancashire, England, which promises “Free Beer Tomorrow,” forever For the past several decades, InfoVis has shared a top spot with artificial intelligence as part of the official (and eternally unreachable) future of humancomputer interaction Don’t get me wrong I’m quite pleased to have AI and InfoVis in the future, provided they don’t get ahead of themselves and try barging into the present, where they don’t belong Let me explain As an information architect, I specify the structure and behavior of web sites, software products, and interactive services so that users can achieve goals, complete tasks, and find what they need In this role, I’m often asked by my clients to evaluate the potential of search engines, content management systems, and other software to support their user experience strategy, and it’s not uncommon in this process for naïve views about InfoVis to surface and cause trouble Inevitably, clients have seen a cool demo—perhaps it was Grokker, Kartoo, Newsmap, The Honeycomb, or TheBrain—and they’re all fired up to transform the user experience This is when I start throwing buckets of cold water on my clients, which is not always fun or politically astute but it is infinitely preferable to designing a disaster Peter Morville is president of Semantic Studios and co-founder of the Information Architecture Institute His books include Information Architecture for the World Wide Web and Ambient Findability, both published by O’Reilly He blogs at http://findability.org To grok or not to grok The problem with InfoVis is that it’s exciting and interesting but not particularly useful for solving most challenges of mainstream information retrieval Yet, because there’s so much money in Web search and enterprise search, software vendors keep spinning InfoVis as a next-generation search solution For instance, Grokker is a “web-based enterprise search management platform that leverages the power of federated content access and visualization to maximize the value of information assets.” Even better, Grokker has a clustering engine that “combines common phrase discovery with latent semantic indexing to extract key concepts.” In other words, Grokker combines AI and InfoVis to automagically categorize and visually represent search results It all sounds great, until you try it, and discover that neither the categories nor the visual maps actually help you find anything Even Grokker now defaults to the Outline View, a rough approximation of the highly successful Guided Navigation model pioneered by Endeca, which relies on hybrid human-computer solutions to taxonomy and metadata development, content categorization, and search Grokker doesn’t grok Its visual display of categories is not informed by nor conducive to profound understanding And with respect to findability, Grokker fails on multiple levels, and so all the other contenders in this category : 19 Release 2.0.4 August 2007 Still Tomorrow’s Technology Peter Morville Grokker fails on multiple levels First, the automatically-generated categories (e.g., Data Visualization, Web, General, Group, More) are sufficiently haphazard to be worse than nothing Second, the spatial rendering of document and cluster icons within larger clusters is harder to understand and scan than a traditional text outline Also, Grokker adds unnecessary steps and complexity to the user’s search process Unknown unknowns That’s not to say that InfoVis has no future To the contrary, in concert with text analytics solutions from such companies as Attensity and Nstein, visualization is already being successfully applied to the challenges of national security and business intelligence By identifying the semantic patterns that emerge from unstructured data streams, these solutions help us to uncover what Donald Rumsfeld labeled the “unknown unknowns.” InfoVis can help us find anomalies and trends we didn’t know to seek However, we must be wary of seductive charts that use past performance to forecast future results We should heed the wisdom of Warren Buffett who “realized technical analysis didn’t work when I turned the charts upside down and didn’t get a different answer” and Peter Lynch who noted “charts are great for predicting the past.” Using vision to think So, how we describe the true potential of InfoVis? The best explanation I’ve heard is from Karl Fast, an information architecture and knowledge management researcher at Kent State University Karl argues that interactive visualizations can help us better understand an existing data set and problem domain by shifting the burden from cognitive to perceptual InfoVis is about using vision to think Karl also admits that InfoVis remains a research problem rather than a practical solution That’s why my advice is to hold off on buying any InfoVis software products, especially in the search domain, at least until tomorrow, when you can finally enjoy that free beer nn 20 : David Weinberger Caution: Low Visibility Release 2.0.4 August 2007 Caution: Low Visibility A contrarian look at the way things look by David Weinberger Listen Online I understand the theory behind showing information visually I’ve got it firmly in mind and could describe it to you in a few terse sentences, neatened up with properly placed commas and perhaps a subordinate clause or two The one thing I could not is draw you a diagram There are left brained people and right brained people And then there are the no brained people who can’t make head nor tail out of your clever visualizations We are the graphic challenged, the chart impaired, the illustration phobic For us unlucky few, your scatter chart remains scattered, your pie chart unconsumed, your bar chart disbarred We are the ones sitting in the back of the room, seeming to stare appreciatively at how you’ve managed to get quarterly sales data correlated with the prime rate, the periodicity of movies about inspiring but doomed teachers, and the movement of corpuscles through hardened arteries, all expressed as a set of dental X-rays At least that’s what it looks like to us Our brows are furrowed not because we’re trying to glean every drop of wisdom from these correlations but because we can’t figure it out We thought we had the gist of it until you announced that the base isn’t zero, it’s 190, the scale is logarithmic, and the up axis—is that the X or the Y?—measures different things on the left than on the right There are always more ways to discover that we misinterpreted it The key you so thoughtfully included is for us just a checklist of disappointment, informing us element by element that our initial understanding was way off It’s there we learn that we can’t even count on the one constant that’s enabled us to fake our way through hundreds of strategic presentations, for in your precious chart, up and to the right is where the company doesn’t want to be Your intentions are good, but your premises are flawed You think that the more information you pack into a visualization, the more revelatory it will be For example, take Edward Tufte’s favorite illustration of all time: The map of Napoleon’s army’s approach to and retreat from Moscow It shows time, space, the number of troops, the seasons, the color of Napoleon’s horse, and the winning recipe in the Pillsbury Borscht-Off If it packed in any more information, the illustration would achieve sentience and possibly dominate us poor Earthlings But for me and those who suffer from my syndrome, it is at best a map of the route Napoleon’s troops took, obscured by meaningless symbols and gestures It’s as if someone thought she could give you better directions to downtown by singing the national anthem and burping the alphabet at the same time Just tell us what we need to know and don’t distract us with your awesome informational multitasking Please If you want to make yourself clear, put things in a row, using a single point of comparison And then spell it out for us “See how this one is taller than that one? That means we made more money this quarter than last.” Don’t ruin it by then pointing out that the thickness of the columns indicates how you and your spouse are getting along and the length of their shadow represents human mortality One thing at a time And, preferably, use your words nn David Weinberger’s most recent book is Everything Is Miscellaneous His other books include The Cluetrain Manifesto and Small Pieces Loosely Joined He is a fellow at Harvard Law School’s Berkman Center for the Internet and Society He served as the senior Internet adviser to the Howard Dean presidential campaign Find him online at http://www.JohoTheBlog.com : 21 Release 2.0.4 August 2007 Vista Perception Critical Update Jimmy Guterman The Number : Vista Perception Critical Update Each issue we will look at one statistic in business and technology and explore what is happening behind that number Beyond the quips, Microsoft’s new operating system might not be fairing as poorly as some critics claim—and hope by Jimmy Guterman “The whole industry is disappointed with Windows Vista” — Acer president Gianfanco Lanci 22 : How much respect is Vista, the most recent version of Microsoft’s Windows operating system, getting? It’s not just the makers of the “I’m a Mac/I’m a PC” commercials who are having fun at Vista’s expense Let’s ask a less biased observer: Ask a Ninja In a recent episode of his web video, the mysterious warrior/comic offered the ultimate “Yo mama” joke: “Yo mama’s so stupid she bought Windows Vista.” When even fictional ninjas are having fun at the expense of a computer software program, you know it’s in trouble It’s not just in pop culture that Vista has been found wanting Gianfranco Lanci, the president of PC maker Acer, has said “the whole [hardware] industry is disappointed with Windows Vista.” Microsoft CFO Chris Liddell has upped the percentage of Windows licenses the company expects to sell in its next fiscal year that are for XP, its six-year-old model, and lowered its Vista expectations It’s a far cry from the December 2006 report by market researcher International Data Corp., which predicted that the new operating system would generate 100,000 new jobs and $70 billion in new revenue for corporations in the U.S alone by the end of this year Surveys taken recently say that less than 10 percent—perhaps as few as percent—of machines now running Windows are running Vista But is Vista the bust that so many observers seem to hope it is? Depending on whose numbers you believe, between 89 and 93 percent of all PCs on the planet run on Windows Mac OS X, Ubuntu, and lesser alternate operating systems may be making headway, but Microsoft has so much dominance—and so much money in the bank—that it could weather the financial ramifications that would come as a result of Vista adoption slowing dramatically until word of mouth—or the product itself—improves (Indeed, the week I’m writing this, in early August, Microsoft has soft-launched “productivity” and “reliability” upgrade packs for Vista.) Even if we’re entering a period of computing in which the operating system on a client PC does not reflect the cloud-based programs we’re using on a regular basis, it would take many years for an even moderately savvy Microsoft to fall on anything approaching hard times There are signs that maybe Vista—and the notion of operating systems as important—are not busts Although early reports encouraged skepticism, recent research by Roger Magoulas and Ben Lorica of O’Reilly’s Research team suggests it’s way too early to stick a fork into either Vista or its product category Magoulas and Lorica analyzed sales of books about computers, a longtime reliable tool for determining what categories and individual products are on their way up or down, and what they saw surprised them Despite the collective wisdom among alpha geeks that operating systems aren’t so important anymore to consumers, the consumers who buy computer books don’t agree The release of Vista and the continued reader interest in XP, coupled with the large unit sales of Windows books as a group, shows a relative increase in the share of computer books sold that are consumer OS books What book sales reveal about Vista Over time, the Vista sales trend looks more like the pattern for the most recent generation books about Mac OS X (10.4, Tiger), with one big difference Says Magoulas, “Sales for the older version of Mac OS X dropped off much more quickly than sales of XP books are dropping A few explanations come to mind: (1) the large and less engaged mass audience for XP/Vista; (2) Apple’s marketing and pricing that successfully encourage users to upgrade; (3) the widely reported stance by many corporations to slowly convert to Vista; (4) the lack of compelling new features in Vista (with the focus on compelling, I’m sure there are plenty of new features).” Indeed there are many new features, graphical and otherwise, enough of them to suggest that maybe, for once, despite the company’s reputation of delivering products long after competitors, Microsoft’s new operating system came out too early—and not just because of bugs in the software eWeek’s Joe Wilcox reports that at a 2003 developers’ conference, when Microsoft was just rolling up its sleeves on Vista development, Bill Gates said he expected the standard PC configuration in 2006 to include a processor of at least GHz, memory of at least G bytes, and a hard drive with a terabyte of storage But that doesn’t define today’s standard—that’s the very high end With its Aero graphical overlay and sundry modules that are hungry for RAM and disk space, Vista is a software platform built for a mainstream hardware platform that doesn’t exist yet It’s still early in Vista’s life Interest in books about it may be flattening, as books about new operating systems versions tend to, but there are sundry client- and server-based upgrades on the horizon that will lure more to pay attention and maybe buy And even if Microsoft, as promised, takes the Draconian measure next year of taking XP off its product list and selling only Vista, don’t expect an immediate mass defection, particularly on the slow-to-change corporate-buying side Vista is certainly not a big hit out of the gate like Windows 95 That may be because operating systems aren’t as important to consumers as they were in 1995 But Microsoft would have to screw up a lot worse for a long time for Windows to lose its standing as the desktop standard, especially in business We’ll see whether Vista can right itself, or whether its launch portends a continued slide nn Despite high-profile delays, it’s possible that Vista came out too early–and not just because of bugs When it comes to book sales, Vista’s quick rise, fall, and plateau mirrors that of previous operating systems At last check, the sales for books about Office 2003 and Office 2007 seemed about to cross, something Vista books did more quickly vs XP : 23 Release 2.0.4 August 2007 Calendar Calendar T Events Tim O’Reilly plans to attend J Events Jimmy Guterman plans to attend Lack of a symbol is no indication of lack of merit Please contact Jimmy (jimmy@guterman.com) to let us know about other events we should include A selection of significant events over the next few months August 27–September Burning Man (Black Rock City, NV) www.burningman.com Artists, technologists, gawkers: head for the desert! September IgniteBoston (Boston, MA) www.oreillynet.com/ignite/blog/ The parent company of this newsletter hosts a night of launching, hacking, learning, and networking September 8–9 The Singularity Summit (San Francisco, CA) www.singlist.org/summit2007/ Prepare for our eventual future as subjects of our robot and/or computer overlords September 17–18 TechCrunch20 (San Francisco, CA) www.techcrunch20.com The inaugural conference from Web entrepreneurs Michael Arrington and Jason Calacanis September 24–26 DEMOfall (San Diego, CA) www.demo.com Chris Shipley promises to show you the future of the technology business September 25–27 J October 5–6 October 17–19 October 18–21 24 : Emerging Technologies Conference (Cambridge, MA) www.technologyreview.com/events/tretc/ MIT hosts a meeting of entrepreneurs, investors, and observers We’re most looking forward to the keynote presentation by Gilberto Gil, Brazil’s Minister of Culture—and one of the greatest musicians that country has produced CommunityNext:Platform (Palo Alto, CA) www.communitynext.com The latest on social platforms and APIs, featuring a $10,000 code-off T J Web 2.0 Summit (San Francisco, CA) www.web2summit.com Produced by O’Reilly, in conjunction with CMP Come and explore the web’s edge (This will also be the subject of the next issue of Release 2.0.) Pop!Tech (Camden, ME) www.poptech.org Over the course of three days, every part of your mind is engaged And, if not, there’s always the lobster October 30 The Future of Business Media (New York, NY) www.fobmconference.com The paidContent gang will examine how digital media and consolidation have turned business and trade media upside down November 6–8 T November Future Forward (Weston, MA) www.futureforward.com This seventh annual executive retreat brings together the top dogs and cats in New England technology and investing November 15–16 T December 2–4 February 6–7 Web 2.0 Expo Tokyo (Tokyo, Japan) www.cmptech.jp/web2expo/eng After a few days off, the Web 2.0 Expo moves on to Asia EG2007 (Los Angeles, CA) www.the-eg.com Even in alleged retirement, Richard Saul Wurman is still focused on making information entertaining and entertainment informative T February 27–March March 3–6 Web 2.0 Expo Berlin (Berlin, Germany) www.web2expo.com/berlin The expo celebrating our favorite meme goes to Europe T J Money:Tech (New York, NY) conferences.oreilly.com/money Paul Kedrosky and the O’Reilly Radar team will show you how to hack Wall Street T TED (Monterey, CA) www.ted.com Yes, the marquee event celebrating technology, entertainment, and design is long sold out Have you signed up for the 2009 installment yet? J O’Reilly Emerging Technology Conference (San Diego, CA) conferences.oreilly.com/etech/ This year the first two days will overlap with our Emerging Telephony Conference : 25 Release 2.0 Subscription Form Complete this form and join the other industry executives who rely on Release 2.0 to stay ahead of the headlines You can also subscribe online at http://r2.oreilly.com Your annual Release 2.0 subscription costs $495 per year, and includes both the print and electronic versions of six every-other-month issues, access to the complete Release 1.0 and Release 2.0 archives, a discount on attendance to O’Reilly conferences, and full access to our website, http://r2.oreilly.com name title company address city state zip country telephone fax email (personal email required for electronic access) url n My colleagues should read Release 2.0, too! Send me information about multiple copy subscriptions and electronic site licenses n check enclosed n charge my (check one) n american express n mastercard card number expiration date name and billing address n same as above n see below n visa cvs code name address city state zip country signature Please fax this form to 1.818.487.4501 or mail it to Release 2.0, P.O Box 17046, North Hollywood CA 91615-9588 Payment must be included with this form Your satisfaction is guaranteed or your money back If you have any questions, please call us at 1.800.889.8969 or 1.707.827.7019 or email us at r2@oreilly.com