Over the last few days I have been attending a workshop at Princeton University hosted by the Center for Information Technology Policy. I am excited that I had this opportunity to listen to some of the thought leaders in this space, the intersection of technology and public policy, and the implications of technology choices for public policy and also the implications of public policy for technologist. Quite insightful discussions. Be warned. What follows are my inchoate notes on the whole affair. It’s all a jumble since I suck as listening and writing at the same time.
Panel 1: Possession and ownership of data:
Moderator: Ed Felten, Director, CITP
Joel Reidenberg, Professor of Law, Fordham University
Timothy B. Lee, blogger at Technology Liberation Front and adjunct scholar, Cato Institute
Marc Rotenberg, Executive Director, Electronic Privacy Information Center
Tim Lee mentioned recent Facebook/Scoble/Plaxo issue. Scoble’s account was disabled for violation of term of service because he used an automated feature of Plaxo to scrape his data from his Facebook account into his new Plaxo account.
Are cookies still a privacy/security issue?
Gmail – Peopple consider it creepy that Google scan’s email to provide advertising but no one seems bothered that it also does that to provide spam and virus filtering.
Facebook – How is data about me shared within and among facebook contacts? How do I shared one piece of data with A but not with B and vice versa? Facebook beacon ( http://www.facebook.com/business/?beacon )
Joel Reidenberg: Cloud Computing can be used to improve privacy. Location is important. Consent model is important. Technical restraints of system architecture determines how information is managed. Easier to control how data is shared in a centralized system than in a distributed system.
What type of system do you need for Ownership Rights? How do you revoke rights?
Information Asymmetry – once data is given to someone – lost of control.
We understand credentials – yes or no answer – are you who you say you are?
Privacy is about transparency. To be more open and accountable about the data collected. Transparency must be enforced.
Marc disagreeing with Tim Lee.
Thinks Google crosses the line from parsing email to provide value added service to using that data for their own or others gain.
Cookies are opaque. I can’t inspect the information stored within or any say in how the cookie is being used.
Rebuttal from Tim Lee:
Panel 2: Security and risk in the cloud
How does the move to centralized services affect the security and reliability of users’ interactions with technology? What new threats are likely to emerge? How might provider behavior, user behavior, or government policy need to change in response to those threats? How does the “open source” ethos work in a cloud computing environment?
Moderator: Alex Halderman, CITP
Marc Hedlund, founder and chief product officer, Wesabe.com
Mihai Christodorescu (home page), researcher at IBM TJ Watson Research Center
Benjamin Mako Hill — researcher at MIT Media Lab and Free Software Foundation
Are the ways in which we categorize data correct? Perhaps not.
Are the ways in which we categorize data correct?
One stop shopping for government and courts?
No good use for scripting in a word processor. How to handle malicious code embedded in trusted sites.
Mako Hill, control over our lives via technology. Free software versus Open Source software. Give user control over how software interacts with their systems.
How do we improve security in the cloud? Cab we expect the industry providers to respond to the need? Do we need policy maker involvement? What is the level of liability ( if any ) required? Do we need full disclosure?
Day 2 of Computing in the Cloud Workshop.
Panel 3: Civics in the cloud How and where can cloud computing best improve public knowledge and engagement in political issues? What has been achieved so far? What is possible in the long run? What moves by private actors, and what policy changes, might do the most to harness the power of cloud computing for civic engagement?
Moderator: David Robinson — Associate Director, CITP
Josh Tauberer (home page), founder of Govtrack.us
Andrew Page, associate director, MAPLight.org
John Wonderlich, Program Director, Sunlight Foundation
Using DNS SNooping to build a profile of the sites that user visit. Can track domains but not individual pages in site.
Tracking hit rates at political candidate web sites. Some of the data is skewed due to software running on sites recursively. After adjusting and looking at trend lines we see that obama’s traffic is trending up while clinton’s is trending down.
Multi-linear regression analysis – Looking at how candidate web sites respond to predictor variables from blogs. e.g dailykos.com, politico.com, economist.com. None of the predictors are good for Obama’s site. Unexplained.
Josh Tauberer – Government should provide access to data in a standard open format (e.g XML ).
Semantic web. Publish data in RDF.
Use data to follow the money – Andrew Page
Panel 4: What’s next?
What new services might develop, and how will today’s services evolve? How well will cloud computing be likely to serve users, companies, investors, government, and the public over the longer run? Which social and policy problems will get worse due to cloud computing, and which will get better?
Moderator: Andrea LaPaugh (home page) — Professor of Computer Science, Princeton University
Reihan Salam — The Atlantic Monthly
Jesse Robbins — O’Reilly Radar
Jonathan Rochelle — Google
Jonathan was impressed with one laptop per child. Most everything we do is moving to the Cloud. Value for customers and service providers. Trend: Collaboration.
Concept of micro-innovation. Value in great content. Differentiation between unique content and commoditized content. Service integration is continuing trend – mash-ups. E.g. Google Maps with Craig’s List. A new supply chain creating platforms for creativity. Do it your self tools and languages are gaining in popularity.
Drive to collaborate brings about awareness that there is no obvious ownership of things that are created. Group can collaborate on a document/service and one person can at a minute lock out all collaborators.
Jesse Robbins – People are using services without an explicit understanding of the service they are using. Can we create service level agreements for free services ( like Gmail or Google Maps )?
Move toward utility computing. How many of want to build our own data centers in out basement? It’s now cool anymore. Leverage Amazon EC2, S3, Google etc. to build out services.
Danger of lock in. Federation will play a role as smaller players band together to offer aggregated services.
Small innovators eventually become large players due to network effect and then become protectionist.
Conversation needs to change between technologist, academics and policy makers. Frequency and honest partnership between all stake holders.
Reihan Salam – Thinks that new applications will be developed that we can’t even predict. Thinks we are thinking too small about the Computing in the Cloud. Mentions 23andme project.
Links outsourcing to growth of Baumol’s Cost Disease.
Baumol’s cost disease (also known as the Baumol Effect) is a phenomenon described by William J. Baumol and William G. Bowen in the 1960s. It involves a rise of salaries in jobs that have experienced no increase of labor productivity in response to rising salaries in other jobs which did experience such labor productivity growth. This goes against the theory in classical economics that wages are always closely tied to labor productivity changes.