Guerilla Informatics

Another code brick in the wall - Technology - theage.com.au

John Markof's Another code brick in the wall pulls together examples of many web 2.0 companies, big and small, that offer a new type of internet-based application integration.

Dear PBS: Free Clifford!

PBS has such a great opportunity to increase its viewership through the Internet. Their community-sponsored model is such a good fit here - I hope they have the courage to put their shows up without DRM encumbrances and show others how successful such a strategy can be.

Our family will gladly sponsor WGBH if they do this.

American tech leadership and economy endangered by "Flat World"

Listening to the "The World is Flat", I thought that the flatteners Friedman described were pretty obvious to anyone who pays attention to modern technology and business trends. But when he goes on to describe the impact they have on American jobs and the American economy, I was surprised to find myself agreeing with his cataclysmic predictions if we don't change many of our immigration, education, and employment policies. Starting with the unquestionable premise that brain power is the basis for our economic growth, he suggests that America needs to quickly adapt to the changing world in the following ways:

In immigration, we are fools not to significantly increase visas for highly skilled, foreign workers. They come here, the stay, they innovate, kaching! The terrorist threat since 9/11 has created bad policies - scrutinize the hell out of the visa candidates if you want, but let them in. Our immigration policies for skilled workers are a type of trade protectionism that is misguided - they are based on the belief that employment is a zero sum game - that americans lose the jobs that the foreigners take. The truth is that skilled foreign workers fuel economic growth that increases jobs. And even if there is some displacement, with the world flattening, these people are going to take your job anyway, even if they're not on US soil. Instead policies need to be directed toward...

Job Training and Job Mobility. Friedman suggests a policy of wage insurance that lets someone get assistance (e.g. 50% of wage loss) if they retrain and transition to a new field that requires that they take less pay for a while. He also brings up the importance of making job benefits mobile...health insurance first, and retirement programs second, so that people aren't forced to stay at companies that don't provide them the training/growth that they need.

Coincidentally, I was glad to see the following article on Cnet about some new Democratic initiatives : broadband access, scholarships for math and science, doubling research spending, alternative fuel initiatives. All these are specifically mentioned by Friedman.

I wonder if the Democrats could really seize on this Flattening theme to create policies that provide a stronger contrast between their polices and those championed by Republicans. Like the old Democratic party that supported the worker along with labor unions, the new Democratic party would support the worker's aspiration to survive and thrive as a worker on a global playing field. Don't have the skills? We'll train you. Worried about your kids ability to compete? We'll fix their schools. Feeling put down by your employer? We'll change the rules so that you can go work somewhere else more easily. The contrast with Republican corporate interest is easy to see and to explain. If people are allowed to change jobs, work for themselves, work flex time, etc., it shifts the balance of power from the corporation to the individual. Many high tech workers already enjoy this type of control....pushing this down the socioeconomic scale would be a big win for Democrats.

IT Conversations: Patrick Grady - Supernova 2005

IT Conversations: Patrick Grady - Supernova 2005: Good historical perspective on attempts to create services centered around the individual. As former VC, Mr. Grady certainly knows how to describe the potential of the market very well. In this talk, he gives one example of users spending an average of 2.5 hours a day attempting to coordinate schedules and make service-oriented purchases. The missing ingredients include knowledge of one's schedule, one's preferences, and secure agents (my term) that can monitor availability of things and buy/book them for you. With the right tools (his Company's), right security, and right scale (he calls "network effect"), he thinks he can succeed where so many have failed thus far (e.g. MS Passport and Haelstorm). He speaks of having a million users of his product, and he makes some good arguments, so I think it's worth a listen.

Living with interruptions, yet still getting things done

Ever wonder what the true impact of interruptions were in your work? We live with "continuous partial attention" reports a human factors researcher at Microsoft featured in this article. You'll also see references to Allen's Book "Getting Things Done" (see link in margin) and his followers such as Danny O'Brian and Merlin Mann at 43 Folders. Another researcher at Microsoft describes his work on having a computer to recognize a user's "busy state" by seeing how much they type, watching them with a webcam and listening to them with a microphone. Ideally, such systems could know when to avoid interrupting you or when you should avoid interrupting others (when you don't have the social cues from sitting in a cube next to someone).

Meet the Life Hackers - New York Times: By CLIVE THOMPSON Published: October 16, 2005
In 2000, Gloria Mark was hired as a professor at the University of California at Irvine. Until then, she was working as a researcher, living a life of comparative peace. She would spend her days in her lab, enjoying the sense of serene focus that comes from immersing yourself for hours at a time in a single project. But when her faculty job began, that all ended. Mark would arrive at her desk in the morning, full of energy and ready to tackle her to-do list - only to suffer an endless stream of interruptions. No sooner had she started one task than a colleague would e-mail her with an urgent request; when she went to work on that, the phone would ring. At the end of the day, she had been so constantly distracted that she would have accomplished o"

Would anonymous forums work for scientists?

The ever-provocative Russell Beattie caught my attention with this post about reconsidering the benefits of forums that allow Anonymity: "

I’ve been obsessing lately about a new concept I noticed a couple weeks ago. I was looking something up online and found this fascinating overview on Wikipedia of the most popular web forum in the world called ‘2 Channel’ or 2ch. It’s a Japanese site which, according to this stats page, gets over 2.5 million posts a day. Wow. The most amazing thing about the site - and the thing that separates it in my mind from just about every other forum I’ve ever seen is that Anonymity isn’t just permitted, it’s encouraged. More on this in a bit.

Forums with no traffic are useless, registration requires far too much activation energy, and its really hard for a small community to bring up the traffic to make forums worthwhile.

Today I was asked to consider contributing to macresearch.org. I'm no expert on community building, but its clear that low-volume registration-only forums won't work for them, either. I wonder if a merit-driven, anonymous forum for scientists would work here? And even if you did get traffic with this approach, could a fledgling community site bear to host content that was impolite to individuals or corporations (e.g. Apple)? It might be worth giving something a try, because in my opinion, low traffic forums like these have more of a negative than positive effect on a website.

predicting what you want next

So Microsoft filed a patent on a machine learning algorithm that predicts what media you'll want to consume next (e.g. pro-active playlists):The Real Deal On Microsoft's Playlist Patent - Forbes.com

Can anyone give me an example of any program that usefully predicts what you want and pre-fetches it for you? It seems like such a huge leap from giving you stuff you asked for in the past (e.g. that document you opened last) to predicting what you want next (e.g. songs from your favorite singer, ready to go). Sure, Amazon can use the hive mind to tell you other stuff you might like, but they're not pushing a trailer down to your computer, just in case you want to view a movie. There's a cost to pre-fetching (bandwidth, configuration effort), so you've got to make it worthwhile. Perhaps there could be a Akamai-type solution that balances the cost of distributed deployment of stuff with the value from accelerated retrieval. Hell, you could use BitTorrent as the protocol, but what type of personalized content would drive this?

There's also the question of competing alternatives. I can subscribe to news/data/media feed for my favorite writer/singer and control it myself, thank you. If people are doing this effectively, then what's left over for predictive algorithms to help you with? The more non-core a pre-fetched item is, the less value it has (when i do want it), and the more likely it is to be junk (when i don't want it).

the latent pain of ms windows

speaking to a colleague the other day about his laptop backup strategy, i first volunteered my approach. i create a full image of my powerbook so that i can recover everything completely (and even bootup from the firewire drive). i do full backups weekly or just before traveling. i also do smaller backups of documents, mail, and such as "incrementals".

my colleague, an apt developer who uses windows os, had a very different strategy. he said he just backs up important files. when i asked him about how he would deal with a catastrophic failure, he said he'd just install all the applications. i was puzzled because this would surely take a long time, and he said that he would appreciate the "opportunity" of doing such a reinstall to get the computer in a cleaner state.

the latent pain here is that windows offers an incredible drag factor to its users. registries get corrupted, things start to break, and people who use these machines - even professional developers - grow increasingly tolerant of this type of thing. some wait for the opportunity to start with a clean computer, and this probably means buying a new one for those without the skills or wherewithall to reinstall everything(while enduring the old broken one for a hopefully short period).

i would wager that you could add six months of life to the average windows computer (the point when the user can't stand it any more) by formatting the drive and starting again (or Degunking, which Vista is addresssing). by comparison, people hang on to their macintoshes to the bitter end, complaining only that it seems slow, and giving in only because they finally *DO* need that piece of software that only runs on mac os x.

another true macintosh benefit: greater longevity with less latent pain.