Two days ago, I posted about unalog. Guess who posted about unalog way back in August of 2004? Yours truly, but since I'm editing in a browser, there's not much of an automated help option out there. Even worse, I had much more context in the first post.

What would it take to ameliorate this problem? Writing tools tied to retrospection tools. Jeff Jarvis likes to chirp about news being fishwrap, so free the archives. But as Steven Berling Johnson points out, good tools make archives quite valuable.

And I still want a nickel for every time the Memex is invoked.

Thought parkin': No not source code management, aggregator source management. In any news aggregator, the management of sources is just as important as the management of the received items, maybe even moreso. Most webfeed aggregators make you manage them explicitly, but tools like Findory completely manage the sources for you, behind the scenes.

Wonder what the spaces inbetween these design end points looks like?

Tim O'Reilly documents O'Reilly and Associates early efforts pioneering advertising on the Web. This is mainly for anecdotal amusement, however it reminds that I'm so old that once upon a time I was actually highly infatuated with GNNPress. GNNPress was like FrontPage except earlier and better, and essentially dead before blogging even got started. The tool get swallowed up when AOL bought GNN and consequently got GNNPress and GNNServer, which PhilG road to glory as AOLServer.

Man I'm old!

Clay Shirky launches a new group blog, You're It focusing on tagging, taxonomy, and folksonomy. Not sure there's much there to cover though.

Probably time to start a Wiki page collecting tagged bookmark applications, so I can include Marcus Campbell's scuttle. Open source, based upon Php and MySQL.

Thought parkin': I wonder if anyone within the HCI community in general and the social navigation subfield in particular, has looked closely at mechanisms for group, permission, and capabilities management. I get the sense that Flickr get's a lot of mileage out of a relatively simple system, but I'm trying to envision the extension to socialized aggregators.

Given that you want to expose portions of your reading behavior to different communities what are decent mechanisms for expressing one's intent? Throwaway approach: start with buddylists and go from there.

Speaking of open source del.icio.us knockoffs, David Chudnov's unalog looks mighty useful. Python, Quixote, and ZODB inside to boot.

Since Joshua, rightly so, doesn't let the source code of del.icio.us out and about, others wanting to implement tagging need to look elsewhere. Gordon Luk's FreeTag is a combo package of Php and MySQL code providing an API and schema for tagging. Don't know if it will scale to the levels of DIU, but it's a start.

Ad hoc tagging strikes me as one of those bits of functionality that has many obviously easy, yet monstrously broken implementations. Sort of like user, group, and permission management capabilities. The more open examples that are out there the better chance developers have of converging on a decent range of techniques.

I enjoyed my first tasting of Jon Garfunkel's Civilities. I like the measured criticism approach, and Civilities has a distinctive visual design (Drupal inside). Might seem a bit formal in tone given today's web writing, but there's good critical thinking in there. I always wondered what happened to Elwyn Jenkins.

Hat tip to Seth Finkelstein

No not the fist but the grease.

Last July, I foresaw a thousand client side page hacks blooming. This was motivated by Adrian Holovaty writing a Mozilla extension to rewrite AllMusic. Adrian's at it again with a compiler for Grease Monkey scripts to pure Mozilla extensions, Python inside no less. That is a moby hack.

Ben Hyde is entertaining himself documenting all the mayhem that Grease Monkey is causing.

Link parkin': Googler Steve Jenson has released atomfeed a Python module for generating Atom syndication feeds, natch.

Yours truly is teaching an experimental intro class on, essentially, history of computing and very basic programming for non-majors. Think how to convince journalism underclassmen that computing, as an intellectual endeavor, is important.

Today I talked a bit about the Wikipedia in lecture. I even pulled up a Britney Spears page and hand waved about how I could vandalize it live and it would be fixed by the end of class. Note that I'm doing this in a pretty big lecture hall with the page projected to about 10 feet high. (Note to Alex H. Big for us is 20+ students. I feel your pain)

Enterprising student with laptop, and wireless access, vetts my claim, hacks Britney's page, and asks me to refresh. "Britney Spears (is in love with Brian Dennis) ...". Much laughter ensues. You get the picture. (For a bunch of technophobes, NU journalism majors are pretty well equipped)

I roll with the punch, laugh at myself and leave Britney up on the big board, and move on with the discussion. Not more than 2 minutes later, Britney's page has been fixed. The collective power of the Wikipedians has been amply demonstrated.

Best live demo I've ever given in my life.

And +1 on Professor D. He's got a sense of humor and he knows what he's talking about.

Just a headsup, but I am actually still alive. Lots of interesting projects going on and consuming time.

By the by, I'm really starting to appreciate Drupal as a logical progression from a blogging tool to an accessible "content management system". I use the scare quotes because I think any blogging tool is a CMS, just with a very focused data model.

For all you media types looking for a sophisticated, open source platform, Drupal is the content management system behind Bryght, BlufftonToday, and UrbanVancouver.com.

Through a very circuitous and serendipitous path I chanced upon Ken Pomeroy's college bball statistics project. I believe the game scores were collected in a distributed fashion, and then made freely available for sports stats wonks to munge. I guess I'm slowly penetrating the community of peer-2-peer, non-baseball stats collectors I dreamed of when I was a regular on USENET's rec.sport.basketball.[pro,college]. If you have any Google chops, you can figure out who I was.

In any event here's my route. Daily Python URL > Bill Mill > game data > Ken Pomeroy > Ken's blog > Ken's stats project.

Ob geek talk. I essentially have two bins of feeds I'm subscribed to. One for monitoring, and one for serendipity. I need to build a tool that I can hand the serendipity bin to for filtering, clustering, analyzing, and visualization, so I can process potential serendipity items faster. The monitoring bin I could use to drive a focused crawler.

Back to the mines!!

Martijn Faassen has a working combination of Python, libxml, and the ElementTree API, called lxml. Fast as hell XML parsing in a Pythonic API. Nice!

With Jaeger apparently going closed source, Adam Sampson's rawdog might be the best contender for an open source webfeed aggregator with Python inside. At least I think it's open source.

Actually, I'll just hedge and say I need to check the licenses on both ends. Still it's good to know of another starting point for aggregator hacking.

Thinking out loud. Suppose you wanted to experiment with a P2P version of a centralized service, especially a popular design like del.icio.us or flickr. Maybe you can't stand the uptime of a feed indexing service or want to experiment with some extensions to a social bookmark service. Trust me, this is all completely hypothetical. What's the design and implementation process look like? What guarantees can you make and what do you have to give up? I think, and this is mostly speculation, the majority of the academic P2P work probably relies on at least a few reliable, resilient, super nodes providing a backbone.

Alternatively, how far can you push towards P2P purity in implementing one of these popular, large scale Web services?

I'm sure someone somewhere must have hashed this out, but is there anything worthwhile in having buddylists for webfeed aggregators? For IM they provide presence and awareness. For aggregators they might provide the basis of a permission management scheme for distribution of subscription and readership information. Oh and if the aggregator supported per item tagging that might be made available to folks as well. Users could allow their aggregator to "chat" with other useful services and agents out there on the 'Net.

So previously I was despairing of annotated images ever catching on, which nukes my 2005 as "Year of FotoNotes" prediction. Matt Haughey may have saved my bacon, with his "memory map" concept.

Take one Web based, high res, satellite image service, with an elegant user interface. Mix with social photo service. Stir in some sentimental geographically based remembrances. Add in a dash of A-List credibility and connectivity.

Voila! The Memory Map social space on Flickr.

Still, someone needs to implement hyperlinks within the annotations for some really explosive effects.

Robin Good gathers together in one place a number of useful royalty free stock photo sites.

I could probably recommend just about every Jay Rosen post, but this recent mixed bag of commentary is especially worthwhile because it touches a couple of times on citizen's journalism. In particular, I found the reaction of some Canadians to the new effort Dose, interesting. Even though the Web supports it, folks will be suspicious of importing the infrastructure for citizen's media. At least if you spew a lot of rhetoric about how connected you are to the communities you purport to serve, but are based in a foreign land. Looks like a case study in the making.

Blogrunner demonstrates the ends to which watch engines can be put, in monitoring what the blogospheres are saying about The Gray Lady. Tools like these are increasingly within the horsepower of low end computers and the storage is chump change. Someone's going to come along and shrink wrap this stuff, or turn it into an elegant hosted web service, and all heck is going to break loose.

Technically it's just an aggregator, so there's no reason you couldn't pull similar tricks for individuals.

Apropos of nothing, I found Doug Rushkoff's first column for Arthur magazine quite inspiring. I travel in some circles where collective action gets short shrift . If enabling and exploring new forms of people peacefully getting stuff done together is a hallmark of our era, times aren't too bad.

Thought parkin. One of the few metrics for evaluating content on the Web seems to be popularity. Easy to calculate, relatively, and seems to correlate with quality for at least "some" community. This is partially why folks love stuff like Technorati and Blogdex. (Is Blogdex even still alive? Does anyone care?).

What if you're evaluating a site, post, podcast, whatever, and you have no popularity information. Are there any other clues that might be indicators of quality?

One that comes to mind is consistency/regularity/reliability, which is why I liked the calenders that used to be on most blogs. You could get a quick metric of at least how long and frequently the author had been doing their thing. Skill improves with repetition so you'd presume that the longer you write/blog, the better you get at it. Plus it's an indication of commitment, which is an initial condition for quality.

Finally longevity and consistency is a metric a machine process could easily determine and might be something worthwhile for content recommenders to add in to recommended items.

Steve Yelvington, one of the principals at Morris Digital Works, announces the launch of BlufftonToday.com. BlufftonToday is another citizen's journalism driven effort. While not ridiculously new, it's been going on in Vancouver (among other places) for a while, this might be portending a trend. Guys like Yelvington have major legitimacy within the news industry, validation from him, Dan Gillmor, Jay Rosen et. al. give other folks cover. Mavericks and innovators at news organizations won't have such a tough row to hoe.

Don't get me wrong, it's really early and unclear how the business angle will work out. However, an environment of considered experimentation is a good thing.

Besides, I just like pointing at stuff that's influenced, however tangentially, by goskokie.com.

Just the executive summary is a gutbomb to digest, but the Kaiser Family Foundation's report on Generation M: Media in the Lives of 8-18 Year Olds is chock full of interesting points. Based on a survey of over 2,000 youth the report is essentially a study of how, and with what media, young folks spend their time.

TV and music are dominant, although computers and the Internet have caught up with console video games in usage time. And the penetration of these devices is over 50% across all strata of society.

The interesting takeaways for me? Videogames and PCs don't displace TV and music. Time spent with media refuses to go down. Kids just do more multitasking. TV and IM. Homework and file sharing.

Reading the newspaper happened for 34% of the sample. The average time spent with the newspaper was only 6 minutes per day, although I can't tell if that's across the entire sample (bad) or those who actually read the paper (worse). Yikes!!

Via Gloria Pan @ morph.

The contents are all old news to me, but Merrill Brown's report to the Carnegie Corporation of New York on how young folks are ditching the concept of traditional news is a complete package for the busy media wonk. Besides, the Carnegie name gives it gravitas, so it's at least a good citation for future media oriented papers and grant proposals.

I agreed with just about everything in Brown's report, but one thing rang really hollow:

Despite these innovations, some experts still warn that the news business

Back in September, I postulated that Web apps needed extensibility, but that you can't pull the old plug-in, or embedded scripting language tricks. What's a model for extending GMail?

Jon Udell may have named the solution to the conundrum. Instead of the app being extensible, it can have an architecture of intermediation. Udell is a bit sketchy on the details, but I'm guessing it involves generating protocol responses in structured formats that support transfomation. Also, the protocol can be proxied and manipulated at levels higher than dealing with HTTP requests and responses.

Simon Willison puts two and two together and identifies Greasemonkey as enabling the browser to become an agent of intermediation. With a little help from the server end, essentially well structured documents, your browser can start to do a heck of a lot more than render the results for you. Heck, it can even start talking to other distributed services, allowing you to extend things like del.icio.us without having to wait on jschacter.

Looks like Joshua Schachter is going to go full time on del.icio.us development. Implications unclear. Let the buyout speculation be..er continue.

Link parkin': CNet summarizes Andrew Odlyzko and Benjamin Tilly's demonstration that all subgroups aren't equally valuable in a network. In fact, bad actors make some subsets have negative value, e.g. spammers.

Ben Hyde chimes in on the topic as well.

Using some agent based simulations from a 2003 Nature paper, Chris Allen speculates that functional human groups actually have an expected size much smaller than the Dunbar number, said number being 150. I'm not sure about the validity of the conclusions, but was interested in the means of modeling and simulating group behavior and coordination. Seemed like a cool thing to hack up in a multi-agent simulator, or even just a good old scripting language.

Of course, some folks believe the Dunbar number is really wrong.

GUESS is Eytan Adar's (HP Labs) toolkit for doing graph exploration and visualization. GUESS is a combination of a custom Jython for querying graphs, JUNG for graph algorithms, and Piccolo, Prefuse, or Touchgraph for visualization. Not quite a Pajek or UCINET replacement but that's the long term goal. The big win is probably the embedded, domain specific, interactive language. And since Adar and his colleagues are working network analysts GUESS is being developed where the rubber meets the road.

Tagging + aggregating seems to be catching on. Michael Heilemann launches aggregato.

I like my personal Findory blogs feed. Serendipitously, I run across some interesting posts from feeds I'm not subscribed to. However, I don't need Findory to tell me about new posts to plasticbag for example. I'm already subscribed with Bloglines.

Of course there doesn't seem to be a good reason why my aggregator couldn't tell Findory that, other than that there isn't any way to do so.

Take this request with a grain of salt though. First, I'm probably an outlier power user who uses Findory and an aggregator, whereas Findory is aiming to be the primary aggregator for many folks. Second, providing a well designed, well performing API is not without challenges. Although as Linden points out, for some tasks so what if they're hard, the're worth doing anyway.

A while ago, I talked about how the Daily Me is here and that a key aspect would be focusing on personal information streams.

The eminent David Strom backs me up.

From the vantage of someone who's been watching online media for over 10 years, Nora Paul takes a look back at some of the anticipated benefits. Examining the potential for the Web to let reporter's communicate better with the audience, she has this quote:

The harshest reality that news organizations have to face is that readers are finding each other, cutting out the

While Joshua is dicking around with barcharts, the del.icio.us knockoffs exploring really interesting functionality are coming.

Here's one vote for just getting the dang inbox working.

Persuadio's interactive blog link exploration site MyDensity exemplifies horrible infoviz execution, when it actually works. I only point to it because I know my favorite visualiation toolkit prefuse, is inside, and MyDensity doesn't even bother to give Jeff Heer's fine work credit.

Link parkin': Seb Sauvage's webgobbler remixes digital images. Python inside.

I sort of admire the folks working hard on social designs for television viewing, like Tom Coates' thoughtful prototype renderings. They're really, really trying. Admirably sisyphean.

Unfortunately, this all seems so oxymoronic, because TV viewing is fundamentally anti-social. Anything designed, both in form and content, to dominate your visual cognitive capacity doesn't leave many spare cycles for recognizing interaction opportunities with your pals, much less higher order social interaction.

There's a reason it's called an Idiot Box.