I came across a very nice visualization by Fairyring (Stephanie Hendrick, a student at the HumLab) after finding that she had been experimenting with TAPoR visualizations and posting them to her Flickr account. In her comments she said she had used the visualization system, GUESS: The Graph Exploration System.
Peter O pointed me to a new phenomenon on the web that I’ve been meaning to blog for a while. That is the leveraging of human players for tasks that can’t be easily automated. Perhaps the best example is the ESP Game. The online game is described in “How to Play”:
The ESP Game is a two-player game. Each time you play you are randomly paired with another player whose identity you don’t know. You can’t communicate with your partner, and the only thing you have in common with them is that you can both see the same image. The goal is to guess what your partner is typing on each image. Once you both type the same word(s), you get a new image.
The game (and its Google Image Labeler spin-off) leverages fun to get image tagging done. Remember when we thought computer image recognition would do that? Now we are using online games to make it fun for humans to do what we do best – instant complex judgements about the visual. If you get enough people playing we could make serious inroads into tagging the visual web.
What is impressive about ESP is what a simple and powerful idea it is and this is Luis von Ahn‘s second sweet contribution, the first one being CAPTCHA and reCAPTCHA.
While it isn’t quite as clean, a generalized version of the idea of people power is Amazon’s Mechanical Turk. The idea is that people can,
Complete simple tasks that people do better than computers. And, get paid for it. Learn more.
Choose from thousands of tasks, control when you work, and decide how much you earn.
Developers can register tasks, people can work on HITs (Human Intelligence Tasks) and get paid for the work, and Amazon can become the largest labour market for small tasks.
netzspannung.org is a German new media group with an archive of “media art, projects from IT research, and lectures on media theory as well as on aesthetics and art history.” They have a number of interfaces to this archive, for an explanation see, Archive Interfaces. The most interesting is the Java Semantic Map (see picture above.)
netzspannung.org is an Internet platform for artistic production, media projects, and intermedia research. As an interface between media art, media technology and society, it functions as an information pool for artists, designers, computer scientists and cultural scientists. Headed by » Monika Fleischmann and » Wolfgang Strauss, at the » MARS Exploratory Media Lab, interdisciplinary teams of architects, artists, designers, computer scientists, art and media scientists are developing and producing tools and interfaces, artistic projects and events at the interface between art and research. All developments and productions are realised in the context of national and international projects.
See The Semantic Map Interface for more on their Java Web Start archive browser.
Humanist posted an announcement for a new issue of the Brazilian journal Text Digital that includes some interesting animated experiments (like the image above) including a series a-writing by Gerard Dalmon. The address “To the reader” starts with,
To weave, write and inscribe thoughts on the digital medium is the purpose of this journal that reaches its fifth number with a somewhat different content. It is the first time we publish an issue with more creative than theoretic interventions.
Shawn sent me this link for the The Most Unusual Books of the World. Loyal readers will have seen my Text in the Machine experiment on Flickr (where there is a photoessay).
McMaster’s archives actually have a number of English fore-edge painted books that were, apparently, popular gifts in their time.
I’m trying to imagine a visualization tool that would show you selected passages cut sculpturally out of a 3D book.
At the beginning of November I was asked to give a keynote for a Digital Scholarship/Digital Libraries symposium at the beautiful of Emory Conference Centre. My talk was titled “The Social Text: Mashing Electronic Texts and Tools” and my thesis was that we needed to forge a closer relationship between scholarly projects and digital libraries. This is a two-fold call for change:
A copy of the slides in PDF is up for FTP access. The file is 15 MB.
We Feel Fine is “an exploration of human emotion, in six movements” that harvests recent blog posts for “I feel” or “I am feeling” and then stores information for visualization. There is an applet where you can look for “men who are abiding”. Thanks to Guy for this.
Visuwords online graphical dictionary and thesaurus is a tool that visualizes WordNet relationships. It shows the synonyms and definitions for words you search for. Thanks to Shawn for this.
Two days ago, on the day of All Hallows (All Saints), Google announced OpenSocial a collection of APIs for embedded social applications. Actually much of the online documentation like the first OpenSocial API Blog entry didn’t go up until early in the morning on November 2nd after the Campfire talk. On November 1st they had their rather hokey Campfire One in one of the open spaces in the Googleplex. A sort of Halloween for older boys.
Screen from YouTube video. Note the campfire monitors.
OpenSocial, is however important to tool development in the humanities. It provides an open model for the type of energetic development we saw in the summer after the Facebook Platform was launched. If it proves rich enough, it will provide a way digital libraries and online e-text sites can open their interface to research tools developed in the community. It could allow us tool developers to create tools that can easily be added by researchers to their sites – tools that are social and can draw on remote sources of data to mashup with the local text. This could enable an open mashup of information that is at the heart of research. It also gives libraries a way to let in tools like the TAPoR Tool bar. For that matter we might see creative tools coming from out students as they fiddle with the technology in ways we can’t imagine.
The key difference between OpenSocial and the Facebook Platform is that the latter is limited to social applications for Facebook, as brilliant as it is. OpenSocial can be used by any host container or social app builder. Some of the other host sites that have committed to using is are Ning and Slide. Speaking of Ning, Marc Andreessen has the best explanations of the significance of both the Facebook Platform phenomenon and OpenSocial potential in his blog, blog.pmarca.com (gander the other stuff on Ning and OpenSocial too).
The New York Times has created another neat text visualization, this time for the Republican Debate. The visualization has two panels. One shows the video, a transcript, and sections. You can jump the video using the transcript or section outline. The other is a “Transcript Analyzer” where you can see a rich prospect of the debate divided by speeches and you can search for words. What is missing is some sort of overview of what the high frequency words are and how they collocate.
So, I have created a public text for analysis in TAPoR and here are some results. Here is a list of words that are high frequency generated using the List Words tool. Some interesting words:
People (76), Think (66), Know (48), Giuliani (42), Clinton (33), Reagan (13), Democrats (16), Republicans (11)
Health (45), Government (35), Security (35), Country (25), Policy (16), Military (15), School (15),
Marriage (23), Insurance (23), Conservative (23), Private (22), Let (21), Gay (12)
Iraq (13), Iran (12), Turkey (7), Canada (2), Darn (2), Europe (5),
Immigrants (5), Citizens (2)
Man (7), Mean (7), Woman (4), Congressman (25)
Answer (10), Problem (10), Solution (5), War (12)
Continue reading Republican Debate: Analyzing the Details – The New York Times