This is just for grins. Mark Newman has cooked up a network based, ranking algorithm for college football teams, with all the mathematical rigor you might expect of a world class physicist.
I wonder how well it predicts future performance given past results?
I've noted here and seen elsewhere usage of Sleepycat's implementation of Berkeley DB for persistent storage. I've started really pushing on an application that uses it as well, and I have to see it's a really handy tool.
I'm on the board of directors for a scrappy little, youth technology/media arts organization called Street Level Youth Media. Today was the annual benefit, a very fine affair that also reminded me that a lot of the board has turned over recently and after 10 years the organization is looking to a new chapter. Street Level is interesting in that it's one of the few places in Chicago where underserved young people can get a healthy dose of video and audio production training.
Which reminded me of David Weinberger's post about Digital Bicycle. In essence, Digital Bicycle is about building peer-to-peer technologies for grass roots media organizations to distribute their works. And since the software is going to be open source, looks perfect for the thrifty non-profit organization. Sounds right up Street Level's alley.
I'm pretty sure that like me, Jack Vinson uses his aggregator as a knowlege foraging tool, but he's way more demanding than I am. Check out his list of features in an ideal feed aggregator. Then again he earns his bread in knowledge management.
One interesting theme that has been lost through the marginalization of Radio Userland is that space of features oriented around writing from your aggregator. NetNewsWire + MarsEdit is pretty good at this, although the coupling doesn't strike me as tight. I don't know of any other desktop aggregators that attempt this, although I confess I haven't been kicking tires recently.
The low end of aggregators has been commoditized by Yahoo!, MS, Google, et.al., but there's plenty of room for innovation and niche action.
Recently I wondered where the intro/retrospection tools for weblogs were. Jeffrey Veen of Adaptive Path recently sneak peeked Measure Map. Measure Map looks more like a higher level Web statistics package, but at least it advances the agenda.
I'll be interested to see how it actually focuses on blogs. Does it just look at log files, with some special information about the blog? Or does Measure Map watch RSS and comment feeds, then selectively interrogate log data?
Steve Palmer seems to be the leader/lead developer for Vienna 2, an open source RSS aggregator, for OS X, with the spit and polish of NetNewsWire. Given MacOS's inherent scriptability, this could be a good platform to work on exploring a hyper-extensible desktop aggregator. Hey, Brent Simmons is a great guy, and there are other extensible aggregators out there, read Awasu, but occassionally you just don't want to wait for the main developers.
It doesn't come through here that often but I'm a closet sports junkie, and well the last 72 hours have been a bit eventful.
I mean the NHL is back, and you gotta get with that NBA preseason!!
Just kidding. But
Yowsa! Manuel Lima, creator of Blogviz, is collecting various complex network visualization projects at: VisualCompelxity.com. There's already over 150 entries, many with quite stunning thumbnail screen captures.
I may just be hallucinating, but Firefox 1.5 Beta seems a heck of a lot faster. Hopefully they put the dang browser on a memory diet. We'll see if it falls apart under my typical onslaught of open tabs.
Meanwhile, Simon Willison neatly summarizes some of the interesting functionality changes upcoming.
Brian Benzinger, at Solution Watch, collected links to a number of tools which visualize del.icio.us behavior, either for an individual or globally. Handy to have them all in one place.
infosthetics recently pointed out that CNet has strarted putting up network visualizations of its story space. Looks like the viz is built off of the affiliation network of terms and stories.
Now there's a network infoviz researcher's wet dreams of data. I'd put my money on a) most users not using the feature, and b) of those that do, a high percentage getting confused and only using it a couple of times. But I'm just the cranky contrarian. However, CNet has a big enough dataset of information and users to draw some real conclusions, with some rigorous analysis.
Here's a feature I'd kill for in these visualizations. They give zero help in maintaining trails through the information landscape. At least provide a shopping cart, fancy it up however you need to for eye candy cred maintenance, of waypoints so I can keep track of the interesting stuff I run across.
Update: Apologies for the typo in Andrew's name.
Serendipity in action. Thanks to reading Nelson Minar, I now know you can navigate tabs in Firefox using Ctrl-PgUp and Ctrl-PgDn. Hands not leaving the keyboard when editing, even in a Web browser, is a good thing.
Link parkin': Kent Bye working out the math on generating the HTML for tagclouds.
Just thinking out loud. Where's the tool that will watch what I write on my blog then analyze, summarize, and help remix it?. Inspired by thinking about a Movable Type into del.icio.us plug-in, but being convinced that's probably not the right place to put that functionality since editing becomes reliant on the uptime of another service.
The idea is that you write a post in MT, using a custom syntax mark some of the links as destined for del.icio.us, sit back and relax. Think of it as an alternative, and familiar, DIU posting interface.
Tim O'Reilly edition.
Lucas Gonze, who's measured, steady work on Webjay I find highly impressive, is saying good things about Ning. Maybe I'll have to take a closer look.
I was checking in on Eszter Hargittai, a colleague in NU's School of Communications, to see if her group had published some results related to political weblogs, only to found out she's launched a whole group, the Web Use Project, and attendant blog.
I've mentioned before that the TSB program in the School of Communications is making some interesting things happen around here, and Eszter's a big part of that.
P.S. This year's TSB speaker series, is looking as good as last year's, including Bruno Latour coming to town. All talks open to the public.
Self-admittedly late to the Web 2.0 party, Scott Berkun nevertheless manages to ask some really good questions, regarding social biases in the new generation of Web applications. Early adopters (can) drive the initial feature push, while latecomers bring different persepctives and desires. Inevitable friction? Similarly, can such systems tolerate high volume users and intermittent participants?
Interestingly, there's actually probably a decent body of academic and empirical work to build on, given the age and attraction of USENET, MOOs, MUDs, and MMORPGs. The question is, what are the contrasting features of Web 2.0 apps with those hoary old vets.
My guess? The Web underpinnings of Web 2.0 apps lead to much looser coupling of individuals and, on average, less frequent interactions. Net effect is you've got hordes of acquintances in a large shared environment. Maybe the correct analogy is that of a mega rock concert or festival.
Regular readers will note that the issue of social boundaries is a recurring theme here at NMH, not that I've actually done much about it. Berkun's questions were refreshing in that they didn't focus on RSS, AJAX, mashing up, or other buzzwords of the day.
Greg Linden watches Google Kirkland so I don't have to.
Jeff Dean of Google is giving a talk at UW on Oct 18th about BigTable, a system for noodling around with petabytes (!!) of data spread across thousands (!!) of machines. Thankfully it's on a Tuesday when my schedule is mostly clear and I can catch the webcast.
Brent Simmons' wonderful NetNewsWire, has been bought by NewsGator. Kewl!! I always call NetNewsWire the gold standard of aggregators, and now that I'm back on the Mac at home, I'm really enjoying it's elegance once again.
NewsGator is now the only (?) webfeed aggregator vendor with a product on both major desktops, in Outlook, and has a web version. If they made a big deal with a blog/feed search engine, that would be a formidable client and service backend combination. Alternatively, NewsGator has to look tastier to folks like AskJeeves or Rupert Murdoch.
The blogosphere has gone bananas over the Ning playground. From my quick scan of the developer docs and FAQS, we're talking about a hosted PHP app service, with some extra support for the fancy widgets of the day (tagclouds), webfeeds as sources and sinks, user/group management, and the ability to "clone" applications. There's no inherent support, that I can tell, for social applications, other than connections to many other Web apps which folks throw in the social bin.
Here's a challenge for someone who wants to build a social application platform. Come up with a calculus of social roles and boundaries. Design and implement a developer's toolkit that makes it easy for application writers to build social environments along a spectrum of concerns and expose social controls to their end users.
I'm surprised at how many folks have fallen for the buzzword compliance. Then again I have been known to be wrong.
Earlier this year, I mentioned the FeedTree project out of Rice University. Dan Sandler was nice enough, simply on that mention, to clue me in to the alpha release of the project. Due to various and sundry events throughout August and September, I didn't have a chance to follow up, but I'll plug them again since the project appears to be quite a bit past alpha now.
As a reminder, the idea is to use P2P mechanisms to distribute webfeed changes, presumably helping publishers deal with hordes of bandwidth hungry aggregators. As an added bonus, updates can be pushed to clients before a regularly scheduled pull, leading to more timely updates.
Despite Tim Bray's extolling the virtues of Post and Poll, FeedTree is a worthy experiment. So if you've got an extra aggregator lying around, or publish a popular feed, look into giving FeedTree a try. What have you got to lose?
Actually the credit goes to Brad Parker, but Bill Clementson has the scoop on MIT finally giving up the rights to the LISPM source code for the greater good. +5 yourself on the geek scale if you understand why this is way cool and are now raising a toast to Stallman, Greenblatt, and Moon.
Remember Budd's little speech about sword's in Kill Bill Vol 2.? The LISPM was a Hattori Hanzo.
Via Ben Hyde
Given that the O'Reilly organization is latching onto the term, and Matt Webb has significantly expounded upon it, attentuation will be the next hot buzzword. Looking at the strict dictionary definition, I'm not sure that's actually a positive thing , but I believe they're focusing more on the engineering sense: reduction of signal amplitude to focus attention, which is a good thing.
While I think Webb's piece could benefit from a bit of attenuation itself, it usefully highlights how a number of mechanisms we use on real life to manage huge flows of information, haven't migrated into our Web based media. Thus there are huge design and implementation opportunities and challenges still to be exploited.
One of the attenuation/locality mechanisms that I've been pondering a bit is the power of temporal organization. The good old calendar and timeline have been dismissed recently in the weblog world as being non-functional, but I think there's a second breath of life for them. If small calendars become a bit more interactive (read AJAXy), the social norms and understandings we bring to them are pretty powerful.
For the most part, you don't have to teach people how to navigate a calendar, and a well designed, compact, interface like Bederson's DateLens provides a nice gateway to any number of underlying data icebergs. Out of the box it's a focus+context inteface using time as an attenuation device. In addition, people immediately bring their own landmarks (birthdays, deadlines, holidays, conferences, anniversaries) to the landscape.
Lots of mileage still to be traveled.
Link parkin': John Musser's programmableweb monitor's public Web service APIs and catalogs them nicely.
Rollyo has been all over the blogosphere, but I'm going to pile on because it's a damn good idea. As Rael Dornfest puts it, Rollyo supports attenuated search. Using the power of the Yahoo! Search API, you can focus searches on specific sets of sites.
Here's a no brainer, make it easy for people to import blogrolls as searchrolls. Voila! You can do focused searches against sources you already trust to see what they have to say on a topic.
Besides, any site with the luscious Rosario Dawson on board gets double thumbs up from me.
Sports edition:
Consider the source, but the Computing Research Association breaks down the shape of the US Govt's Fiscal Year 2006 research appropriations. The skinny? Things were looking grim pre-Katrina, with the real dollar equivalent of budget cuts being applied to many of the agencies (NSF, DOE, NASA, et. al.) where academic computing gets its funding. Post-Katrina, the odds are looking up for even deeper across the board cuts at places like NSF.
With the Department of Defense trending towards more short term, technology transferish projects, funding is getting hairy for research computing. Boo hoo for ivory tower types I know, but computing used to be a national competitive advantage. Now we're falling back to the pack.
Sports, biting wit, blogging. What's not to like about Gawker media's new blog Deadspin? Anybody who thinks Woody Paige is a buffoon gets bonus points in my book.
Thanks to l.m. orchard, I now know that XBox Media Center is eminently hackable, even having some nice Python based tools onboard. XBox Media Center is an open source media player that runs on top of MicroSoft's XBox. In essence, you get a platform for turning your XBox into a hackable home media player.
This might be the excuse I need to actually get a game console. Also, you could probably hook kids who don't have a lot of money/wherewithal for a desktop (think disadvantaged urban youth), into some programming.
flagrantdisregard, a.k.a. John Watson, has been noodling about with the Flickr API for awhile now, and has accreted a fine set of toys that create things from photos posted on Flickr.
Apropos of a whimsical Friday evening, last week was New Student Week at NU, and by a set of weird circumstances I wound up having to help advise freshmen. The original colleague on duty had a conference to attend. His substitute suddently got called away to do a DARPA dog and pony show, leaving moi' to be part of the first impression 17 or so youngsters received from Northwestern.
No big deal other than this was a reminder that 20 years ago, I was a freshman at M.I.T. There's nothing like interacting with a bunch of folks who didn't even exist when you were a freshman in college to bring your mortality home.
To commemorate this savage reminder from above, I am now christening a new (the only?) regular feature of New Media Hack, "I'm So Old..." scheduled to appear on Fridays. In this week's edition:
Misery loves company, that is all.
Innovation in the blog publishing tool space has been a bit slow lately if you ask me. Self tagging for tools like Technorati, and a re-emergence of link blogging, energized by Web based tools like del.icio.us are about the extent of the excitement. Correct me if I'm wrong oh loyal readers.
SixApart, which had a big hand in launching a couple of the last Big Things (TM) is making noises as if it wants to strike gold a second time with Planet Comet, detailed a bit more in this press release and by Richard Macmanus on his Web 2.0 ZDNet blog.
From the screen captures though, I'm hard pressed to see anything really spanking new. A lot of the photo sharing sites, e.g. Fotki are starting to morph and integrate much of what I can see of Project Comet.
Findory, headed up by Greg Linden, just unleashed a Bloglines style feed aggregator, but with Findory's personalization mechanisms baked in. Explicit focus plus manufactured serendipity. Will definitely have to kick the tires on this one.
Kevin Burton, a founder of Rojo recently set sail for new shores. His new venture is TailRank which is in that terminally annoying invite-only phase (thanks for reminding me to forget about your product until later), but still seems interesting. Looks like TailRank will be a weblog and webfeed search engine that will try to use social information to present results further from the popular peaks.
Not to pick on Kevin, but if Web 2.0 crashes and burns spectacularly, I'm sure invite-only betas will be a highly mocked archetype of the era.
No smilies for the humour impaired.
Apropos of nothing, while it wasn't too painful getting back on the air, it feels a bit trickier getting back in the posting groove. I was cooking throughout the summer. Let's see if I can't close out the year in fine fashion.
In an earlier post, I conjectured that laptops are "the briefcase of the new millenium". I'm glad to see I'm in good company as none other than Peter Merholz wonders why product designers don't take advantage of laptop mobility. He's also another observer of the growing trend towards the laptop showing up in leisure time, probably due to its capabilities as a coordination (synching with others), navigation (maps, directions), and interstitial (dvds, games, photo munging) tool.
Guilty as charged for taking the laptop on vacation.
If I didn't know any better I'd think Porcupine == (Zope - HTTP server), given the feature list.
Just what the world needs. Yet Another Python Web Application Framework.
Howard Rheingold ran a class at Stanford entitled "The Literacy of Cooperation". The course focused on how humans collectively get things done, outside of well studied "market forces". There were a number of interesting speakers for the class, and their sessions were recorded. I'm particularly interested in what Bernardo Huberman and Paul Hartzog had to say.
Sort of. The new DNS entries for the NU CS, errr EECS, department are taking forevah to propagate. So costarica's hardware is on the air but its name to ip address mapping is a bit dodgy at the moment.
Like I said, there'd be some rough edges.