Andrew Jaffe: Leaves on the Line

Welcome to anyone one led here from Physics World’s Blog life column. This is a blog — so comments are encouraged (or you could click on the advertisements)!

There’s a new Google-competing search engine called Cuil (which I guess is meant to be pronounced as slacker-speak “kewl” or something). If I search for myself on it, my Imperial homepage comes up first, but for some reason accompanied by this picture. I promise that’s not me. Just as strange, a picture that is of me comes up next to a blurb for a book written by another Andrew Jaffe who happens to be the director of the Clio Awards for advertising (there are a few of us Andrew Jaffes out there, but I’m egotistically happy to tell you that I’m top in most search engines). But I’m happiest about the cover shot from Muscle Magazine (not me either).

Over the last few weeks, I’ve managed to get hold of invitations to a few semi-private (semi-public?) beta versions of interesting bits of software. OK, they’re not for Royal Ascot or even a posh dinner party, but I have to take what I can get. So be prepared for some serious geekery — sorry.

Evernote wants to be your “offline brain” — a place where you can store all of the interesting detritus of your life that you might want to recall later: this “place” is on the internet, in “the cloud” as people are starting to say, accessible from your desktop, your mobile phone, etc. The further cleverness comes that this isn’t just a place to store files, but a whole lot of so-called metadata besides. So Evernote knows the type of files that you give it, when they were uploaded, and any “tags” or descriptions you care to attach. But the cleverest bit of all is that it can generate its own metadata: it can read text in your files. Well, that’s the idea. I know the camera on my iPhone isn’t very good, but it’s a surprising that it can’t read “Moro” on the left, although it gets “Against” on the right (note that the stored versions are better than these tiny reproductions):

Right now, the software is a web-based service (including one specialized for mobile devices like the iPhone), and offline clients for Mac OSX and Windows. Eventually, they claim it will move beyond character recognition to faces, video, and anything else that can be automatically parsed. Well, someday.

Still, even if they haven’t quite got it working yet, the visionary ideas behind Evernote are best described in this interview with Evernote CEO (?) Phil Libin by John Udell (whose excellent ongoing podcast series “Interviews with Innovators” treads an interesting path between the web and academia).

I’ve also managed to get an account on the Microsoft Mesh “tech preview”, discussed all over the web last week, and, in particular, touted at length on this Gillmor Gang podcast (which is how I got my invitation). I admit, I don’t quite get it — or at least it’s pretty clear that we’re in very early days here. Apparently, the Mesh will be a repository for “all” of your data, translated as needed to different formats and synchronized with different devices, with granular control of who gets to see it and use it. I guess the point is that this becomes the basis for a net-based operating system (perhaps more so when combined with something like MS’s Silverlight to handle the user interface—these all Steve Gillmor’s ideas, not my own).

But at present it’s not an improvement over (say) Apple’s .Mac service, which lets me have an “iDisk” which I can access from anywhere and which shows up on my OSX desktop as a local disk, and which furthermore automatically synchronizes my calendar, contacts, etc. But I’ll wait until the Mac and iPhone clients come around before I pass judgement. And of course there’s an opening for it to be combined with the intelligence of a system like Evernote.

Apparently, I’ve got 10 invitations for the Evernote beta-test and five for Mesh: get in touch if you’re interested. I can, of course, be bribed with, say, co-authorships (I’m an academic, after all).

Google has just released a new version of its Google Earth software — one that lets you look up to the sky instead of down to the ground. It’s essentially a consumer-grade Virtual Observatory, like the UK AstroGrid, the US National Virtual Observatory and the Euro-VO project.

It’s not so obvious when you fire it up and are presented with little icons for various stars and galaxies, but the underlying data is a continuous picture of the sky, although the resolution depends on what data exists in a given area. For example, type in “HDF” and it takes you to the Hubble Deep Field North, one of the deepest images ever taken of the sky, showing galaxies in every stage of their evolution. Conversely, however, most of the objects don’t have any information attached to them at all — just fuzzy blobs.

Of course, real astronomers would require a lot more information: how was the data taken? At what frequency? It would certainly be great to be able to use this as a front-end to the “real” Virtual Observatory like AstroGrid. These science-oriented projects have spent a considerable amount of time and effort refining their interface, but just don’t have the funds or expertise of a company like Google. And now I’m just waiting for someone to implement a layer showing the Cosmic Microwave Background and other “diffuse” sets of data on the sky. (Update: my very bright grad student, JZ, has figured out how to import CMB data as an image into the program.)

Update 2: VO/Blogger Alasdair Allan has started to work out how to connect Google Sky to the Virtual Observatory via the PLASTIC protocol. Alasdair was also interviewed about Google Sky for the Guardian’s science podcast.

Update 3: See the comments for information on Sky-Map, a web-based project very similar to Google Sky, as well as a pretty complete dataset showing the microwave sky.

One of the tenets of the so-called Web 2.0 is that it’s about an “architecture of participation”, allowing users (i.e., everyone) to contribute their knowledge and expertise — or just enthusiasm — to harness our “collective intelligence”. That’s why Wikipedia is about as good as the Britannica — and why you can look up photos of pretty much anything out there on Flickr.

But now we’ll be able to put that collective intelligence to some really good use: astronomy. One of the hardest, most human-intensive tasks has always been “morphological classification”, that is, using how something looks to assign it to a particular group. In particular, astronomers since the early 20th Century have classified galaxies according to categories first enumerated by Edwin Hubble (who also discovered the expansion of the Universe that underlies the Big Bang). Galaxies are either Spirals (like all the famous pictures that you have seen) or Ellipticals (round, smoothly-distributed agglomerations of stars) with subcategories describing the details (and at least some of the underlying physics).

For perfect images, this is the kind of problems that Artificial Intelligence techniques like neural networks have been able to handle with some success. But with realistic, messy data humans are still much better — and if lots of people can vote, it’s even possible to iron out disagreements. At GalaxyZoo, a group of young astronomers, including Oxford’s Kate Land, Chris Lintott, Anse Slosar, and Kevin Schawinski, have teamed up with the Sloan Digital Sky Survey, the largest single dataset of galaxy images yet produced, and created a website that lets the rest of us out on the internet help with that classification — no formal training required, although there are practice images and a test to make sure you can handle the pressures of modern science.

My mobile rang this afternoon, but as usual at work, I missed the call. I didn’t recognize the number (02920368701 in the UK), but on a whim I googled it. Lo and behold, it’s a cold-call mobile spam number. Good thing I didn’t answer it, after all…

My favorite small museum is the Tenement Museum down in New York’s Lower East Side. It was built in the late 19th Century, housing a steady stream of immigrants until it was boarded up in the middle of the 20th. Different apartments in the building have been recreated as they might have been during different decades of the building’s history, and as it was left in 1935.

At first, the tenement housed Irish Catholics, then, European Ashkenazi and Sephardic Jews, and, finally, Italian Catholics. There is a list of the building’s tenants, and when I visited I found that the last names of all of my grandparents — the common Jewish names of Levy, Cohen, Greenberg and Jaffe — were all represented. There’s no evidence that any of my relations ever lived there, but the stories are familiar: escaping persecution in the old world and searching for success in the new one, a short stay in a tenement, before moving onto a more stable existence in an established immigrant community (for example, Pelham Parkway in the Bronx in my father’s parents case). And then, for their children, a move to the suburbs to raise the next generation — including me.

But the Tenement Museum has a larger presence than just its physical location. The immigrants haven’t stopped coming, and neither has the museum. It sponsors the Digital Artist in Residence Project (DARP). The wonderful and fun Folk Songs for the Five Points collects music and found sounds from the still-immigrant neighborhood and lets you mix and match the sounds to create contemporary folk songs, and soak yourself in the noises of the Lower East Side in 2007.

The blogosphere is in echo-chamber mode with advice and admonition for how bloggers should behave (don’t be rude), and how scientists should spread the rational word (don’t be boring).

We take so much of the web for granted today, we often forget how very contingent it all is. Without the very specific work by Tim Berners-Lee inventing the http protocol, perhaps some sort of hypertext communication standard would have come along, but it’s hard to believe that it would be quite the same. Berners-Lee has always advocated a still more open “read/write” web, and about the closest we come to that is, of course, the weblog. Well, blogs were arguably launched ten years ago, on April 1, 1997, by Dave Winer. Scripting News was an outgrowth of his DaveNet emails, but had all the usual hallmarks of a blog: short items, lots of links, and, crucially, reverse chronological order. Dave has gone onto a career as a general computer pundit and curmudgeon — and also invented RSS (that orange “XML Feed” icon over at the side).

Of course, the first of April has another name. If you read this blog regularly, you know that we physicists have a fantastic collective sense of humor, as evidenced here, here and here. Funny, huh?

My own April Fool’s incident came a bit early, last Thursday night, unable to make my way from the arrival hall of Rome’s Fiumicino airport to the airport Hilton. I arrived around midnight, after the trains stop running into town. What they don’t tell you is that all the passageways between the airport buildings are also shut — without signs to tell you where to go. After conflicting information from three different sets of people, I found myself staggering around the deserted parking lots searching for the warm bed I had booked (I did eventually find it, and the front desk took pity on me in the form of an upgrade to the “executive suite” floor). The next day, although a bit sleepy, was at least a productive discussion of the next step in our proposal for a new mission to measure the polarization of the microwave background — in about 2015 or 2020.

But since today is really April 2, you can also read a real blog post by Amedeo Balbi, my cosmology colleague (on MAXIMA and Planck and probably more in the future) over in Tommaso Dorigo’s blog; he’s got one of his own, but it will only make sense if you read Italian.

Yesterday evening I attended the launch party for Nature Network London, a new site run by Nature magazine, which hopes to be a web home for science and scientists in London. There are articles, blogs, discussion forums and calendars of scientific events.

Perhaps unsurprisingly, I ended up meeting lots of people from Imperial — whom of course I had never met here on campus. I also met the site’s editor, Matt Brown, as well as blogger Jennifer Rohn, who also runs the science/culture site LabLit.

It’s an ambitious idea, and anything that gets us out of our offices and talking with other scientists is welcome. The formal barriers to entry are quite low, but to get working scientists to spend their time blogging, posting in discussion forums, and just taking this newfangled social web 2.0 thing seriously may be a hard sell. We’ll have to hook ‘em young. However, “science” in London is dominated by Medicine and biology — we physical scientists are a distinct minority, and our interests, academic lives and ways of working are often very different indeed (for example, the biologists last night spent a lot of time trying to decide whether to approach someone like Paul Smith for a design of a fashionable lab coat — I’ve never worn a lab coat in my life!). Anyway, if you’re a London-based scientist of any stripe reading this, sign up and join in!

Tonight I’m off on a 24-hour jaunt to Rome to discuss our proposal for a new Satellite, BPol, to measure the CMB polarization (and thereby discover if inflation could be responsible for getting our Universe into the shape we find it today). Unfortunately, this satellite wouldn’t be launched until the late 2010s, which means that the data wouldn’t flow for a staggering decade and a half.

Luckily, cosmology will remain interesting while we’re waiting — as Tommaso Dorigo’s ongoing reports from our Outstanding questions for the standard cosmological model meeting continue to attest.