All right, I'm going to be off a few days on that hiatus estimate. Our department is being moved, wholesale, to a new building, right at the beginning of the academic year. Servers, including the one supporting this fine blog and webfeed, are going to be shipped willy nilly across campus. This is alleged to start on Wednesday, which means good old costarica will be powered down, sometime on Tuesday. While I can't predict exactly what they are, I'm anticipating some glitches somewhere.

With any luck, I'll be back on the air by Monday. Consider anything else gravy.

Ciao!!

Yeah, I know it's not a word. But between the long weekend, travel with intermittent network access, the upcoming Fall quarter at Northwestern, and to top it all off, the post-Katrina nightmare, it seems like a good time for this here blog to take an extended break. I'll see you on the other side of September 8th with more new media hackery.

Ob link: The fine folks at Flickr have released a major extension to their web service API that let's you find out what photosets, and more importantly, group pools a photo participates in. This is majorly geeky, but you can now take the identifier of a Flickr photo and do some extensive social network navigation using the photo as a starting point. Interestingly, you can walk along human to human connections, and media to media connections, through tags, sets, pools, and contacts.

As promised, for BlogDay, here are five new blogs I've used to seed my Rojo account, first to expand my reading horizon, second to get some experience using Rojo.

Koranteng's Toli is the rambling musings of Koranteng Ofosu-Amaah, a native of Ghana. Ofosu-Amaah is working for IBM/Lotus in my old digs of Cambridge, MA.

Sokari Ekine's Afrotecnik is a technology oriented adjunct to the well known Black Looks.

In an alternate universe, I wouldn't mind being Keith Jenkins, author of Good Reputation Sleeping. DC based photojournalist, now photo editor for The Washington Post, and his son and I apparently share a birthday

Ndesanjo Macha writes about politics, technology, culture, and media on the African continent, in two languages no less!

I find Lynne d Johnson's diary a little overdesigned for my taste, but can't resist the combination of hip-hop and technology. But that's what aggregators are for!! Besides, she's got serious chops, proving yet again that journalism combined with technology fearlessness can lead to powerful effects. She probably doesn't need me to prop her up, though.

Side note: I'm mildly chuffed that I couldn't find a single decent house music weblog. Granted I didn't serch that long, but seems like there's a huge opportunity for some ground level reporting on this club scene. Or if I'm just ignorant, someone please fill me in.

Just for Technorati blogday, BlogDay2005.

Leigh Dodds produced a well written survey of REST and social software intersections. I use the term intersection since many of the services described, the usual suspects, implement the REST philosophy less than completely. Dodd's also discusses some best practices for building such systems in a RESTful fashion.

I ran across this article from Lucas Gonze's detailed examination of Webjay and REST. It's a good look at one developer's real world set of design choices.

Nicholas Chim's Personal Bee project is some confirmation that a lot of what I've been thinking about in terms of aggregators isn't out in left field. The Bee watches bunches of feeds, does some term analysis of the items that come across a bunch, and generates a tagcloud for navigating the items. TechCrunch has a good overview, but it's something like a personal scale Google News.

Lots of mileage in this idea.

Maybe I should institute a a weekly Grand Cayman mention here on NMH.

Hurrican Katrina is a Category 5 hurricane and about to lay waste to a big chunk of the Third Coast.

According to Wikipedia, 2004's Hurricane Ivan was a Cat 5 'cane, when it passed by Grand Cayman. Amazing the level of detail on Wikipedia. When I was on Grand Cayman this past July, 10 months after Ivan, the place was still devastated. The hotel my wife and I stayed at, literally got a brand new beach out of the deal. We were regaled with stories of winds blowing so strong through staircase openings that the air felt like knives. Granted, Grand Cayman isn't the US, but it's not exactly third world.

Just a guess, but this year's Sugar Bowl might be in jeopardy. Heck, Mardi Gras is not a definite.

StoryCast looks like a neat little system for creating and publishing audio annotated photo slideshows with a handheld device. There's a little more detail in an HP techreport.

TechCrunch is a recently launched blog dedicated to coverage of Web 2.0 applications. Edited by Michael Arrington and Keith Teare, TechCrunch is based out of Archimedes Ventures, a venture capital firm. Looks like they're essentially making a bit of their internal research visible to the outside world to garner mindshare and pick up bits of early feedback on emerging technologies. Interesting strategy.

As an example, they've gathered some initial data on Flock, apparently a Firefox mutant with social bookmarking and weblog editing baked in. From what I've seen so far, the TechCrunch team seems to have pretty good radar, so it's probably worth a webfeed subscription.

When Flickr first announced "interestingness" I declared "let the gaming begin". There's at least one Flickr group, with a decent burst of activity, devoted to cataloging interesting photos along with occasional debates on how the ranking algorithm works. Not to mention a few long, interesting, discussion threads, regarding the relative merits of interestingness and discovering good new images.

The more active and/or older members of Flickr seem relatively hostile to interestingness which seems to have a heavy component of popularity baked into it. The more experienced folks seem to want tools for navigating smaller social spheres, as exemplified by this conversation. I'm often struck by how popularity is valuable because it's a cheap, easy indicator, built out of lots of evaluations, of certain qualities. However, once you get out past those low hanging fruit, or start measuring different qualities, the useful indicators are hard to find.

As social media environments become massively richer in, and expose more, metadata it'll be interesting to see how tools for semionauts are designed and constructed. Marcos Weskamp's FlickrGraph is about the closest thing I've seen and that only let's you explore the underlying social network, leaving out photos, the central media of Flickr!!

Just thinking outloud. Lot's of folks are unhappy with Technorati. Technorati has two goals which seem to be giving it a lot of difficulty: comprehensiveness and timeliness. They're trying real hard to watch all of the blogosphere as fast as possible. The scale issue, especially with the explosion of spam blogs, puts severe constraints on the choice of indexing and search algorithms they can use. The speed issue sets up unrealistic expectations, which frankly, I'm not sure add much value. The pre-eminence of the scoop mentality is just a transference from mainstream journalism and doesn't really work on the Web.

What if Technorati developed and promoted another smaller, slower archive? Slower meaning the index is only updated say once a day, or even once a week. Smaller meaning it has a much higher, but transparent, threshold of what constitutes a "blog" and what gets indexed. With the added slack, more sophisticated algorithms for eliminating spam blogs could be incorporated. Similarly, interesting query and navigation tools, like Adina Levine's conversation clouds, could be built. Sure they wouldn't be realtime but sometimes it's good to slow down. Also, given the temporal nature of blogs, useful retrospection tools could be built.

As an added exemplar, Google News severely compromises on the scale axis, but in return provides other useful functionality, e.g. topic clustering.

Movable Type 3.2 is officially out of beta. The free version for individuals has returned, once again with unlimited weblogs.

Steve Lohr of the NY Times does a drive-by on battling dropping computer science enrollment. Two NU connections. First, NU's own Robin Hunicke, told a similar story to Microsoft's games folks earlier this summer. Second, the Jamika Burge mentioned in Lohr's article is a sometimes collaborator of my advisee Azzari Jarrett. I know you're happy for me.

My two cents on the cs enrollment "crisis"? Smart undergrads (and intelligent people in general) have figuread out that there's a lot of interesting and powerful things you can do with computers without having to suffer the rather unpleasant beatdown that Architecture, Operating Systems, Compilers, Networking, and Theory (typical required elements of a CS degree) often inflict. Along with the overwhelming caveman hacker ethic (me like command line), it'll be interesting to see if traditional CS departments can successfully find a way to let go of a few of these pillars and still maintain their engineering and scientific self-respect.

Alternatively, they could go all aggro and start selling hard to people interested in computational biology, nanotech, complex systems and networks, etc. Meanwhile, the "soft" stuff could be left to departments like Berkeley's SIMS, Indiana's School of Informatics, Irvine's school of Information and Computer Sciences, and of course The Media Lab.

The first thing I thought about when I heard that Google Talk was going to be Jabber/XMPP based was, "Wonder if it works with Gush?".

Answer: Affirmative, although the current results are probably a bit disappointing. Also, you don't get the voice stuff that Google's desktop app provides.

Why bother? Well Gush is still an amazingly visually pleasing way to do IM, which counts big if you're going to IM a lot.

Link parkin': Aric Hagberg's NetworkX is a Python package for experimenting with complex networks. Looks like it also incorporates connections with drawing packages such as matplotlib and GraphViz.

Narendra Rocherolle and Nick Wilder, a couple of the hackers behind the long in the tooth Webshots, cooked up Shoebox, a "vertical" social bookmarking and aggregation application. Shoebox is for bookmarking images on the Web, and watching webfeed streams of photos. Pursuant to its focus on images, Shoebox has a decidedly different presentation style. I think there are any number of applications like Shoebox, which ride on top of the webfeed ecology, but to the average user, don't look anything like an aggregator.

Shoebox development has a blog and Rocherolle's contains a lot of insight into Webshots, which is way older and way bigger than Flickr.

Memo to Webshots. Flash popunders are exceedingly painful. Please make it stop!!

Hmmm, with this new set of APIs, it looks like NewsGator is starting to emerge as an interesting platform to develop aggregator experiments. The combination of an online component, a desktop component in FeedDemon, and automated control through the APIs makes a tempting target.

Glad to know that the physics community has taken on the critical issue of proper mathematical treatment of American collegiate football rankings. And that network analysis is an important element!! Only Mark Newman could pull off something like that.

Actually, for those aspiring blogosphere index builders, much of Newman's work should be required reading.

If the Konfabulator crew hadn't merged with Yahoo! they'd be tearing their hair out. Google, while beefing up the Google Desktop app, has added Google Sidebar: "a panel on your desktop which provides convenient, one-glance access to all sorts of personalized information." Said Sidebar has a developer's API to leverage the long tail of free coder time.

A quick scan of the Google Desktop dev docs indicates all sorts of COM interfacing bits, with nasty, pointy sharp teeth. +1 for Konfabulator.

Kurt Cagle, a prominent SVG developer, put together a detailed and philosophical argument about RSS/Atom as a payload for Web services. I noodled about on this topic a few days ago, but Cagle beat me by a month and brings much more depth. Of particular note is the contrast with what's going on in the SOAP community, which seems to be hamstrung by its RPC model and overengineered specifications.

As evidenced by activity on the del.icio.us blog, steam is picking up on feature development for the social bookmark service. Now you can make queries againt your bookmarks and receive the results in JavaScript code or JSON, language neutral JavaScript objects. This will make it easier to incorporate del.icio.us into AJAX, browser side JavaScript, applications since the apps won't have to internalize XML.

Also, this potentially provides a standard client side object model for tagged bookmarks across all JavaScript del.icio.us applications. And conceivably any other bookmark service (Furl, spurl, unalog) could output data in the same format, leading to an "industry" standard.

Link parkin': Joe Gregorio has cooked up a Python module to make URL fetching using ETag and gzip caching easier. Probably should have thrown in If-Modified for extra style points.

Alan Kay gets downsized. NU CS gets contracted. Bell Labs groundbreaking UNIX department is officially dead.

Okay, one of these things doesn't belong but still, moment of silence for the real pid 1... Thank you very much.

Hat tip to Paul Boutin.

I'm about due for another serious big dive into a new programming language (c.f. Norvig "Teach Yourself Programming in 10 Years"). For a while I thought it was going to be Php, but I think I'll just hold my nose and learn just enough to do the Web hacking I need to do. However, Ruby is starting to look a bit more palatable, especially after Ian Bicking's comprehensive feature comparison of Ruby and Python. Ruby, Python "Power" is not really a versus piece with a "scorecard" and "in the final analysis, the winner is" bit, but a well organized work of comparison and contrast.

Yup. It already happened. But as Lilia Efimova points out, with all the noise around Top N lists, rankings, conversations, and clouds it might be a good time to re/visit what's been going on in the academy regarding analyzing the blogosphere, especially in finding community. If you're trawling for useful material, some of Andrew Tomkins work is a good place to start.

Rafat Ali distills the Google Library contretemps into its essence: permission or forgiveness? Looks like the publishers aren't the forgiving kind, especially with coin on the table.

Barb Dybwad's list of suggestions for social bookmark services is good stuff. I can't really do it justice with a summary so just go read it. (And I really hate posts that say that.)

Her list points out one of the issues with centralized Web services. Since these aren't pluggable applications, the rate of innovation is gated by the developers who have access to the backend code. Web service APIs are a start to building new applications but from what I've seen, apps built on top of them rarely get folded back into the main service.

Speaking of spurious lists, Sherman's Travel ranked Chicago in the top 10 of cities for art lovers, along with other venerable destinations like London, Berlin, and Florence.

As a Bay Area refugee, I often get asked to compare the two metro areas. I love both but culturallly, Chicago just crushes SF.

Feedster launches a challenger to the Technorati 100. Kudos for making the raw data available, although what little is known about the ranking algorithm still leaves it pretty opaque.

Some of the squawking about who is and isn't on the list reminds me of two things: 1) the definition of a weblog is a bit tricky, 2) some folks need to be on a list, any list, to prop up their own self-esteem.

I'm not with Jeff Jarvis on much, but tools that focused on surfacing small worlds, would be much more useful

Thought parkin': one of the downsides to designing and building a RESTful application is that you have to come up with a document format for your HTTP responses. Then you have to inform all your clients what that format is. And then they have to figure out how to injest your format.

With the proliferation of webfeeds, any one of the many RSS flavors or Atom could fulfill the document format role in a pinch. There still has to be a bit of coordination on how to interpret the items in the feed, but at least you're a step up the abstraction ladder wth much less coupling of server and clients.

If you can't think of anything better, strikes me that Atom or RSS should just be the default.

Apropos of nothing: The Roaming Gnome visited the Cayman Islands. Photos courtesy of John Carleton.

Thanks to my own trip there, I now have a permanent soft spot for the CI.

Neat work done by Scott Golder and Bernardo Huberman on the dynamics and structure of our favorite tagging system, del.icio.us. The actual data set may not be all that representative (it was taken over a weekend and looks only at popular URLs), but the analysis is clear and rigorous. Maybe efforts like this will nudge Joshua Schachter to release a few comprehensive del.icio.us datasets for examination.

Walt Crawford regularly writes, and writes well, on a wide variety of library and technology related issues. His recent issue of Cites & Insights contains a thorough study of the behavior and reach of a large number of library related weblogs. For those who have the time, one way to make a case for reading non A-list blogs are investigations like these.

Software tools that ease small scale analyses of blog communities could go a long way towards expanding horizons.

Fed up with current feed reader limitations, Greg Linden and Findory came up with their own aggregator. I'm sure it will tend towards simplicity over complexity, but it will be interesting to see how the additional information from users will play with Findory's personalization mechanisms.

Dang! Back in late 2004, early 2005 a colleague and I were kicking around ideas forn an online project with a major media concern. One thing that came up was fantasy leagues for folks who don't like sports, e.g. a Red Carpet Fantasy League for spring awards season. Minor details about team composition and scoring to be determined. Granted it's not a uniquely brilliant idea, but I hadn't seen anyone execute it online.

Enter Erica Salmon's Fantasy Fashion League. You draft major designers and scoring is based upon appearances in major fashion rags like Women's Wear Daily and Vogue. They're even planning for next year's Oscars!!

My only suggestion would be to make the leagues and partcipants a little more visible. Half the fun of these things are seeing what's happening in other circles besides your own. A little egoboo and exhibition play can help fuel a community.

Via Everything And Nothing

The Online Journalism Review has been going through some changes. In general, I've been digging OJR more, mainly because they got a decent RSS feed, but the features have been much deeper recently, and the video journalism project was a nice experiment. Editor-in-chief Robert Niles is exploring new directions for the publication, with a healthy mix of optimism and thoughtful consideration about what's worth doing at OJR. With the support of USC's Annenberg School they have a unique soapbox that affords the potential for big impact.

Found the following quote quite amusing though:

Unfortunately, CARR has remained a specialty within journalism, rather than a core skill. Part of this can be attributed to journalists' collective hostility toward math and science. I've been training journalists in basic math for a decade, and in my experience, it is far easier to teach someone with high math aptitude how to report and write a journalism story than it is to teach a typical journalist math.

There's a bit of rhetoric in journalism education about "understanding an increasingly ccomplex world", but if math/statistical skills are a key tool in understanding complexity (much less computing) what does the above quote portend?

Link parkin': Nice overview of PageRank by Ian Rogers. Chock full of examples to boot!!

I'm going to take up the challenge and try to expand my blog reading horizons by participating in BlogDay 2005.

Hat tip to Ben Hyde.

A long time ago, I was a big devotee of Baseball Prospectus, a baseball site driven by the Bill James school of thought: highly valuing rigorous statistical analysis to explain performance in the sport. They always had great stuff, but I lost track of them after they partially went behind a pay wall. BP's stuff also was very professionally done despite most of the authors not being traditional journalists. It was a nice example of productive citizens media without being blogging.

Well despite my trepidations, it looks like BP is thriving. PaidContent even has them as a potential purchase for the suddenly voracious Fox Interactive Media, although a large grain of salt needs to be taken with the report.

Apropos of nothing. Wow! That BP post was my actual first item with real content. Gosh, has it been that long?

Hefty. Mark Glaser's piece on the NY Times merging its online and print newsrooms is a worthwhile, if time consuming, read. The article is both a feature on the event and an interview with Martin Nisenholtz, tops at NYT Digital, and Bill Keller, big kahuna of the print editorial side.

While there's been a bit of Web triumphalism going on, I'm of a mind that this will only be as good as the number of effective bridgers in the mix. Folks like Adrian Holovaty who understand the rhythm and codes of the newsroom but can really hack too.

This isn't unprecedented, MSNBC has done good stuff with a similar mentality as well as a number of smaller print concerns, but it may be the final stamp of legitimacy needed to push along the slow pokes in the hinterlands. "If it's good enough for The Times...". Along with what the Associated Press is doing, we may also see an impetus for journalism educators to more aggressively address how to prep their students for these new newsrooms.

Then again, there's a part of me that wants to establish an over/under on time to split out a business unit to "capitalize on the unique online opportunities the merged newsroom has been missing." Apples to oranges, but we were digging on that AOL/TimeWarner merger once upon a time too.

ActiveState provides professional support for many of the open source scripting languages. They've just launched a company blog.