yelvington.com

In a special edition, Ifra's Newspaper Techniques tries to explain Web 2.0 to publishers and does a remarkably good job making a very complex subject accessible to a general audience in only 26 magazine pages.

I giggled at a transcript of an interview with Tim O'Reilly that cites the "Jengo." Actually, it's Django, a Python development framework named after the famed Roma jazz guitarist Django Reinhardt of le Quintette du Hot Club de France. (Explanation: Adrian Holovaty is a jazz nut.)

Incidentally, O'Reilly has his own explanation of Web 2.0. I think publishers should stick to Ifra's.

Om Malik profiles a company he calls "the American Idol of digital photography." The idea: Get photographers to compete online to get their work printed in a magazine. The site: JPG magazine.

It's human nature to compete for scarce resources. Online space isn't scarce, but print always is scarce. I wonder what use newspapers could make of this principle. Hmm.

The article also has some hints about how JPG's website was built, taking advantage of new Web services such as Amazon.com's S3.

We think of Amazon.com as a retailer, but it's unveiled some powerful new web services that can make it possible for disruptive new Web 2.0 sites to be developed:

Amazon Simple Storage Service (Amazon S3) enables massive online data storage and simple HTTP delivery in a "cloud," in which you pay only for the storage and bandwidth you use. This makes it possible for a video site, for example, to grow rapidly without having to manage the planning, investment and project management processes associated with adding storage arrays.

Amazon Elastic Compute Cloud (Amazon EC2) extends this principle to computing processes. You can configure a virtual server image -- starting with a Fedora Linux base -- that includes all your applications and configuration. To add a server, you simply place an electronic order through a web services API. Absolutely all of the server setup processes from then on are automated. Your servers are up and running within minutes.

Pricing on both is cheap and tied directly to what you use. Need more servers? Start them up. Need fewer? Release the resources to the grid.

The disruption will kick in as entrepreneurs imagine new ways to use this power.

It's not just about running webservers. Think video transcoding. CGI rendering. Engineering computations. A university could configure an on-demand scalable parallel computing farm for scientific modeling. Since it's all done on a machine-hour basis, 100 machines could work for one hour instead of 1 machine for 100 hours. This doesn't apply to all classes of problems, but many new doors are opened.

This afternoon at Morris DigitalWorks, where I do my day job, we're launching a Web 2.0 social filter metasite about pro and collegiate sports. I alluded to this project last week.

FanaticZone.com is a remix of some current cutting-edge ideas combined with a niche topical focus. You'll recognize some of the ideas from Newsvine, Digg, Beta.netscape.com, and from various RSS readers and aggregators.

When you click through, FZ will take you all over the Internet. That's the point of a metasite -- it's not about publishing content, it's about finding content. Wherever it may be, including photos at Flickr and videos at YouTube.

That finding process is powered by human intelligence and automation working together. FZ also supports user conversation and commenting, and some social networking features with more to come. People power matters.

Ah, but hasn't FanaticZone been around for years? Yes, it has. Back in the Internet bubble, Morris built Fanaticzone as a Southeastern Conference site. We pulled the plug on the old site when the Internet bubble collapsed. It's been on automation ever since, displaying statistical data from SportsTicker that's been run through some sophisticated Morris parsing engines.

Times change. Ideas change.

Our real goal with FZ is to learn. It's like a concept car, a tool to help us work through some ideas and understand their implications. It's incomplete and somewhat raw. We probably have a lot of things wrong. That's OK. It will change. Our users will help shape the direction it takes.

The current site is the result of some internal fiddling. The real site will emerge as users begin to poke on it.

This is a six-week project. You heard it right, six weeks. Nik Wilets, who heads the MDX Lab at Morris DigitalWorks, designed it, led the project, and did some of the coding. Stefanie Rodriguez, an intern, did the bulk of coding and integration work. It is built on the Drupal open-source platform using some contributed modules and some custom work. Underneath is the standard LAMP stack.

Six weeks is an important figure. We're learning that fast development of a flawed product is infinitely more valuable in the long run than slow development that aims for perfection.

We're always going to fall short of perfection. It's more important to discover quickly whether we're directionally correct. Discover your mistakes early in the process. Don't fear failure. If you're going to fear something, fear your own hubris.

This is an extraordinarily hard lesson to learn for those of us who come from businesses with defensive cultures. But it's an important one.