Something that saved me and could save you too

So I went home this weekend for an interview and realized that I forgot my hard drive that had a little demo I had put together on it. Rather than freaking out I calmly searched the internet and came across this, www.dropbox.com. Dropbox is a site that let's you upload up to 2 gigabytes free into a folder that creates a link that other people can download from. I had my friend just drag the 1 gig .mov file into there at Ithaca and then downloaded it super fast at my home. Anyone that's trying to move large files around should check it out.

Games with a Purpose or GWAP's

How many tags?

Tagging has become one of the most invisible but significant elements of digital media. Although it is not immediately apparent, when we search on the web, in our own folders or in the public library, we depend 100% on tags and their accuracy to find what we are looking for.

A tag can be described as a non-hierarchical keyword which can identify a piece of information. In this respect it can be considered metadata, since it is information about information.

Now, tags are usually keywords chosen by someone to describe a particular aspect of that information. In the case of an image for example, tags would vary greatly depending on who "tags" the picture. Since this is not a very reliable way to describe anything, various tools have evolved that attempt to remedy this situation by applying crowdsourcing to solve this problem.

Delicious is one example where each individual tags a particular site with as many keywords that might bring it up again in a future search. We all know how easy it was to completely lose sight of a great site among our bookmarks simply because we forgot either how we named the bookmark or the name of the site itself which can be pretty cryptic anyway, no to mention the fact that before Delicious, bookmarks resided in our computer so we could not access them if we were away from it. So now our bookmarks, that old term, reside in the cloud, for everyone to access.

Google, whose main purpose is to make information accessible to everyone (some might disagree), has developed sophisticated algorithms to classify and tag all the information available on the internet and the cloud. However, tagging and classifying images, videos and music has proven to be specially difficult, because of the same problem we face when we ourselves attempt a description.

So what is the solution? depends on the users of course, to do the dirty work. Google Image Labeler is one such tool that has proven very effective. It's simple, fun, game-like minimal interface allows you and a random partner to label images with ever more complex tags, and it gives you "points" depending on the complexity or precision of the label. For any label to work, both you and your partner must submit the same keyword or descriptive phrase.

This idea was originally proposed by Carnegie Mellon's Luis von Ahn, who also developed the infamous and ubiquitous captchas. Originally called the ESP game, which is when I first played it some years ago, it was subsequently licensed by Google. One important difference is that now, if the picture that is presented to you has been tagged before, all those tags are now off-limits, so you have to come up with new ones. This was in response to the fear that tags would become very generic since it is easier to agree on an obvious word, like bird, instead of for example agree on the specific species name.

I have for a long time been intrigued by how we, humans, have suddenly, as the tools became available, become busy bees tagging, describing, sharing information, even as trivial as to what we are doing at the moment, where we are, what we ate etc. (Twitter and FB being the best examples.) My own weird take on it, is that this is the way the "system", the cloud or the emerging consciousness of our information age, gathers all the necessary bits and pieces to achieve the critical mass or the tipping point in order to wake up. A few years ago this thought was simply part of the SF literature. Today is is just a waiting game.

And please, while you wait, tag your posts!

Think before you click

Google's data center, from Harper's magazine, March 2008

This blueprint depicts Google's data center, code named "02 PROJECT" at The Dalles, Oregon.

Google and its nemesis, Microsoft, Yahoo and Ask.com went there lured by the cheapest electricity in North America. And they need it. They are building gigantic server farms on the banks of the Columbia River.

We tend to think of the internet, the virtual worlds, our email and the "paperless society" as somehow a 'cleaner' option than the legacy of our ancestors. We talk of information and the exchange of ideas, as if these concepts reside on the sphere of intangibles that lead us to a gentler, greener world.

People are slowly realizing that driving their cars and SUV's leave a carbon "footprint" for which we will be hold accountable by generations to come. So we telecommute, we email instead of writing a letter or a note (would we write so much dribble if we had to use pen and ink?) and we search or, using the new verb, we google for information, sometimes for its own sake, the new addiction.

Little do we realize that this new industry is as heavy as they come, and like the article in Harper's 1 mentions: "...an energy glutton that is only growing hungrier." Currently Google is estimated to have in the order of a million servers. According to the article: "...the servers require a half-watt in cooling for every watt they use in processing". Some simple math will yield in the order of 103 megawatts for the Dalles plant alone. As a curious note, the Northwest Aluminum smelter across the street form the data center once used 85 megawatts before falling under soaring energy costs.

If you examine the blueprint above, you will see that 18,800 sq. ft. are devoted to the "cooling towers". Perhaps it will become fashionable again to wear the Top Hats of the 1920's to celebrate the new Information Chimneys of the New Millenium?

Can you identify the gentleman in the picture?
C'mon, I'll buy pizza for the first to name him!

And the winner is...Takumi! Choose your toppings (**>

While you are at it, look at the building on the upper right corner and you will see a suggestive label: "Transient Employee Dormitory Building". As with all things digital, employees are simply bits and blips in the flow of the new oil. And the "carbon footprint" is not pretty.

The article mentions that in 2006 American data centers "consumed more power than American televisions" and that in all locations currently active or in planning, both in America and Siberia, Shanghai and Dublin where AT&T, Microsoft and Google are heading in search (pun intended) of even cheaper electricity, most of this energy is produced by the burning of fossil fuels. I gives new meaning to the term "cloud computing"


1 The Harper's article "KEYWORD:EVIL, Google's addiction to cheap electricity", which inspired this post, was written by Ginger Strand, author of Inventing Niagara: Beauty, Power and Lies, to be published this spring by Simon & Schuster