The student group that worked on GoSkokie.com made their report on hyper-local citizen's media available online (big PDF warning). The micro executive summary: "Hyper local citizen's media sites will proliferate whether traditional media organizations participate or not."

Note to interested readers. Medill's New Media Program has now conducted two hyperlocal capstone projects and two youth product projects. The projects have produced real results and have had real industry impact. For example, YourMom, an offering of the Quad Cities Times is a direct result of a capstone project. We're not quite up to the level of Lawrence.com, but we've got some decent expertise in these two areas. If you're a media company or journalism educator and looking for some hired talent, you could do worse than drop Rich Gordon an e-mail.

Disclaimer: I got a small acknowledgement for providing the hardware and some technical assistance in support of GoSkokie.com.

Alex Sirota's FoxyTunes embeds controls for a heap of music players directly into Firefox's user interface.

One area in which Mozilla/Firefox really has it over other browsers is the ability to actually extend the browser. Previously, in pondering Mozilla extensions for client side control of page display, I mused that Mozilla is way easier to extend than IE, but not trivial. Maybe the threshold is a lot lower than I thought. Even if they're not hard to write, they're a darn site easier to install.

Currently, I have my browser kitted out with the TabBrowser, Live HTTP Headers, and WebDevelopers extensions and I'm going to give FoxyTunes a try. And I'm probably an extension lightweight.

It's going to be interesting if Mozilla/Firefox ushers in a new era of extensible browsers. For a while there, toolbars and side bar development was really hot. Still could be in IE land for all I know. But think about potential synergies between in-page applications (e.g. Flickr, GMail) and in-gui extensions. While the browser might never become the MS killing platform of Netscape high times, there's definitely going to be some serious drag in moving to those rich media desktopish apps.

As if Movable Type didn't have enough whizzy plug-ins, Andrew Sutherland's KoalaRainbow provides a tag set that lets you query the MT data store and generate vizualization renderings.

My description probably isn't doing the tool justice, so check out some KoalaRainbow examples.

Part of the problem with video on the Web is that it just isn't all that sociable, particularly due to the fact that it's hard to cite subparts of long presentations. Jon Udell has been a vigorous advocate for making this task easier.

In trying to help out David Ascher, Udell works through how to cite into RealNetworks streams. Not all that bad once you find out the stream's URL.

Thinking about Flickr jogged my memory to check up on Flickr's progenitor, Fotonotes.net. The Fotonotes site looks a tad dormant, but Greg Elin had been threatening to open source his implementation, the protocols, and APIs.

Turns out the action is over on Elin's weblog, where volunteers are being taken to have a crack at his open source Fotonotes release.

I'd love to raise my hand, but I know just enough to know I'm at my project engagement limit. Maybe, I'll try my hand in mid-October.

The fine folks at Flickr have released a new product: Organizr. The key bits are better browsing (time and tag based), photo collections (including custom presentations of photo groups) and users can define their own groups for social photo management.

Oh, and they plug their API right off the bat and taunt folks into doing something better. This in fact may be the biggest part of the announcement.

It's sort of neat to see the folks at Flickr reposition the company and still innovate. Remember when they first started, it was all about using photos to initiate chat. Now it's more about plain old photo management, integration with other Web publishing tools, and social translucence.

Being able to watch tags in del.icio.us is great, but repetition is a killer. Subscribe to the RSS feed for a popular tag and watch as the same link keeps coming over the transom. Overnight, I can have upwards of 60 links added to my inbox, many duplicates from the day before. I need another wrapper feed to collapse these guys!!

Next generation aggregation tools will need to have distillation and summarization facilities baked in. I don't think they have to be able to solve AI hard problems, but the interfaces need to be very good.

Speaking of wrapping feeds, wouldn't it be nice if del.icio.us tag feeds could call home, login as you, and check to see whether you've already saved and tagged a link?

Aggregator as platform again. Maybe I'll have to get one of those Web based aggregators and hack on it in parallel to my bloglines reading.

In reading the Library Groupware paper, I ran across the now defunct OpCit project run by the Open Archives Initiative. OpCit seemed to be a rigorous project to dig into electronic scholarly link analysis using arXiv as grist. Maybe there's some lessons learned in there for the blog search engines and social bookmarks crowd.

In Ariadne, a UK based digital library magazine, Daniel Chudnov and friends hypothesize the need for library groupware, tools that assist scholars in holdings and reference management. Cloaked in terminology I'm quite unfamiliar with, the essence comes to me as that future users will expect libraries to integrate Web authoring tools (weblogs, social bookmarks). The concept of a "link resolver" comes up and as far as I can tell simply refers to an electronic broker to deal with the fact that most major libraries not only have core holdings, but a lot of other complex access relationships. Therefore, where link traversal should actually wind up becomes a difficult question.

The paper does an excellent job of outlining some of the issues, but could have used some more technical meat for my taste. However, it looks like Chudnov has developed an open source del.icio.us knockoff called unalog

Thomas Heller's ctypes module for Python looks like a nice alternative to pre-processing wrappers like SWIG.

Mainly a thin layer on top of the OS facilities for loading dynamic libraries, with ctypes you don't have to go through the rigamarole of writing a specfication file, translating that file to glue code, and then building an extension.

The down side is that you don't get the signature checking safety net that glue code can provide. Ergo, it's much easier to shoot yourself in the foot, in a big way.

Jon Udell is pimping del.icio.us hard.

When I guy with over 1,800 Bloglines subscriptions, not to mention an Infoworld column, pumps the hype on a tool things are looking up.

Memo to Jon. Get rid of the Google search box. That's so 2002

Inspired by a presentation at Hypertext '04, Matt Webb mashes up del.icio.us, Blogger, and Technorati, in a brilliant loosely coupled scheme. The gambit would allow people to easily and freely tag links, ala del.icio.us, in their weblogs. A de facto standard attribute on a tags (maybe img too?) could carry the tags. Then Technorati, since it does the hard work of crawling all those blogs anyway, could stand in for del.icio.us: generating a front page of recently posted links or creating per tag webfeeds. Heck, they could even provide whizzier search than del.icio.us.

Minor note, Webb didn't actually toss Blogger in the mix. I just used it as a proxy for weblog authoring tools. Continuing in that vein, personally I'd throw Waypath at the task before Technorati. 7 days worth of del.icio.us isn' t all that useful.

Honestly, this scheme is just simple and useful enough to catch on. What it needs is a littlte fine tuning (e.g. what's the simplest, cleanest way to add metadata to links) and some evangelism in the form of "here's how to do it in your weblog tool," preferably with an easy to install plug-in. Knock out LiveJournal, MovableType, WordPress, TypePad, and some Java based blog tool, and you're probably 90% of the way home.

It's so crazy it could work!! At least until the tag spammers show up.

I'd like some of what John Battelle is smoking late nights.

Searchstreams are what Battelle term the first class traces of focused searchers. The trail of links, queries, bookmarks, etc. that people generate on their way to actually finding information that they want. Tying back to Bush's Memex, Battelle makes a persuasive argument that embryonic searchstreams are already here, and that they will be a vital mechanism for making sense of information in the future.

I think he's right, but it'll be interesting to see how this plays out. In an ideal world, browsers would expose searchstream creation, editing, management, publication, and aggregation features. Bang! Cambrian explosion.

Web browser advancement is stuck though. So now you see bits and pieces of searchstreams escaping out in the form of Web applications, bookmarklets, address bar shortcuts, toolbars, context menus, et. al. This leads to somewhat haphazard innovation with uneven results.

Then again, maybe the second style is more in keeping with the Web's precepts and has a better chance to succeed.

RasterWeb agrees, del.icio.us is good stuff due to focus of functionality, clean design, an obvious API and no marketing.

In contrast to the other Web bookmarks competitors, del.icio.us smacks of good taste.

In responding to Jason Kottke's noodling on the entire Web as a computing platform, am I canoodling?

In any event, the thought that Google has built a world scale computer (and I like to think that Yahoo and MSN have as well, maybe even IBM), has pretty much stuck. But who's building the Web PC for the rest of us?

With a nod to Manila and Radio Userland, where are the consumer grade platforms that help glue all of those nifty Web services and standardized formats together? As a developer, why do I have to repurpose the LAMP architecture every time I want to build the next Bloglines, or del.icio.us? It's like programming with no OS.

And don't give me that Zope crap. That's like programming on the Amiga.

Cue Dr. Evil voice. Bloglines is reporting that they store and index one hell of a lot of news stories and blog entires.

I demand 100 million postings!!

David Ascher of ActiveState, recently published an extended essay on dynamic languages. The piece was essentially meant as an update of John Ousterhout's classic paper on scripting languages. Ascher attempts to update the discussion in the face of the changing productivity and deployment demands for software development. We can spend more cycles on higher level languages that make programmers more productive developing Web and distributed apps.

Frankly, I found the essay disappointing for two main reasons.

First, programming language designers and implementors have been using the term "dynamic languages" for quite a while now. Granted it was all those old fuddy, duddy Lisp, Smalltalk, and ML hackers, but when the designers of the current popular crowd look to swipe ideas, who do you think they crib from? All Ascher's paper does is muddy the waters in an attempt to get away from a term "scripting" that isn't all that pejorative anymore.

Second, Ascher's definition of dynamic language mandates that the language be open source. This in itself is non-sensical since "open source"ness is a property of language implementations, not languages themselves. There's nothing that prevented Jython from being closed source. And has anyone seen the source for Microsoft's implementation of JavaScript in IE recently?

Open source, especially access to the source code, is a red herring for this discussion. Open access to the language design process is what developers really want. Scheme, Smalltalk, and Common Lisp all advanced without a "primary" open source implementation, although there were some. The key was that developers could experiment in their own implementations with new language features and then profer those as changes to the language.

An argument could be made that having a primary implementation actually hinders a language's progress. There's little chance radical implementation approaches and dependent design features, ala Scheme, can arise in something like Python due to the limitations of the CPython implementation.

Common Lisp will always be a dynamic language to me, no matter what anyone else says.

The following thought was inspired by a spate of bookmarklets and address bar shortcuts that have been recently generated for del.icio.us. The tipper was Ryan Tomayko's del.icio.us address-barlets and more importantly, his argument about muscle memory setting in.

Newspaper organizations should think of constantly providing boomarklets and address bar shortcuts as a service that ties local information to muscle memory. Advantages abound. Builds loyalty, or lockin if you prefer. You can get more information about users out of their bookmarklet choices and usage. Heck they can be custom generated. They're not malware and ridiculously easy to install (bookmarklets) and explain (address bar shortcuts). Developing them isn't even all that hard.

As of now, news organizations are completely ceding this opportunity to Web application developers e.g.: MSN, Yahoo, Google.

The tricky bit is to come up with services that users would want to routinely use. Coupon or classified lookup for comparison purposes anyone? There's gotta be something weather related in there. Arts and entertainment possibly.

Of course, online news sites would have to consider themselves also in the service building business, as opposed to just the content delivery business.

Just a thought.

Damn if it didn't sneak up on me, or maybe I was just oblivious, but del.icio.us now supports search on your own set of bookmarks. And of course, since it's a RESTful API you can do all sorts of useful stuff straight from the address/URL bar in your browser.

Ryan Tomayko has already put together a nice collection of del.icio.us shortcuts for your entertainment.

I'm on my third aggregator now (NetNewsWire, FeedDemon, Bloglines). All three pretty much present items as a big scroll. Despite some people's protestations, one huge pile of news isn't all that different than a bunch of big piles of news.

But the feeds in my blogroll are diverse, so why not more diversity in presenting them? For example, most of my feeds produce on a human timescale and rate; fewer than double digit posts per day. Then there are few firehoses: Engadget, (old) Scoble, Boing Boing. Now there are synthetic information streams like my del.icio.us inbox which can blow real humans out of the water. For the first category, wouldn't it be nice to have retrospective background information collected and constructed, just to provide context? For the last two categories, I'd like summarizations. Whiffs of AI-hard issues come to mind, but I think a little bit of topic clustering and temporal clustering would go a long way. Not to mention a bit of decent infoviz.

Oh yeah, and with photos pushing their way into syndication why not just tile a bunch of thumbnails?

In a current aggregator, as a first cut users could simply specify an interface preference with a reasonable default. Blue sky, next generation aggregators could adapt to the feed.

Martin Wattenberg of IBM Research has executed an interesting array of research projects. To quote him, his work focuses on "visual explorations of culturally significant data." An offhand glance would place it in the infoviz bin, but Wattenberg's efforts intersect with math, art, and collaborative computing.

A Movable Type Gang of Four (Hammersley, Choate, Allen, Haughey) is working on a book, which looks focused on hacking and extending our favorite blogging system.

Interestingly, from what little I can make out of the mock cover, it's not an O'Reilly book.

Despite a dead streak in early summer, and some name competition from Wired News, we here at New Media Hack are still proud to say that we've still got the top spot on Google for "New Media Hack".

No comments, no backtracks, no redesigns, no blogroll.

Just keeping it simple and stupid.

Elise Bauer does yeoman work trying to determine market share for various blogging tools. The methodology has obvious flaws, e.g. conflating Google indexed URLs with individual sites, but the relative shares of the various tools seem qualitatively in line: LiveJournal and Blogger on top, many other well known tools an order of magnitude less penetration.

Well, other than TypePad having twice the share of MovableType. That one seems a bit odd to me given how much longer MT has been around than TP. I'll profer two possibilities. One, WordPress really is putting a dent into MovableType's installs. Two, the numbers are also significantly capturing a usage factor, and the most enthusiastic and consistent group of posters could be on TypePad.

In the comments, Anil Dash asks for suggested improvements. Write a crawler that uses the Google API to generate a seed set, cluster URLs into individual sites, come up with some heuristics to identify the blogging tool, and let it loose!! For bonus points, do some activity analysis to determine if the blog is alive or abandoned.

This one will be kicking around the blogosphere for a while.

del.icio.us seems to be enabling a spate of innovative experiments. extisp.icio.us generates a map of a user's "tagspace". Buzz Andersen cooked up a Mac OS X desktop client for your social bookmarks. Delicious Mind is Ian Oeschger's hack to take your del.icio.us posts and, through a Java applet, turn them into a mind map. Not to mention a number of bookmarklets that provide more information before posting.

del.icio.us's REST API lends itself quite nicely to these types of explorations. The only major issue that I can see is that del.icio.us is pretty HTTP caching unfriendly. Cookies are always set and none of the useful HTTP headers are set. This makes it a tad difficult to write a truly friendly client and for now people are trying to be nice.

I wonder if the other social bookmarking services are seeing any of this type of hackery? Maybe I should just put this blog full time on the social bookmark beat.

Local Projects is a New York based design firm that seems to be specializing in interactive maps. Looks like they've done a number of interesting projects, oriented around allowing people to annotate maps with their personal memories.

In my work on the Encyclopedia of Chicago History Online, I'm trying to spread the meme of letting all comers, through Web based authoring technologies, be able to annotate primary materials, including many of the historical maps that are going to be included. Seems like Local Projects has some lessons learned that might be valuable.

Alex Halavais, University at Buffalo, is routing significant parts of his survey (?) media/communications course through his weblog. I find now that as a scholar, I'm gravitating to issues that are on his syllabus. Meanwhile, I've never gotten within spitting distance of a masscom class. So I'll be essaying to read up and chime in as much as possible.

Jon Udell describes how David Rumsey was a hit at the Open Source Convention. Rumsey is making a large number of rare maps accessible through the Web instead of sending them to a museum.

The part that caught my eye was the use of a browser based image viewer that allows the digitized maps to be annotated with good old URLs. At the end of my CMS class this spring that soon (next 2-3 years) photos would become real first class citizens on the web. You'll be able to distinguish a photo URL from a generic image URL, pull metadata from photo URLs, and be able to manipulate photos through REST style APIs. Of course you can substitute digital image for photo, but the consumer camera market combined with explosive camphone adoption makes this a no brainer.

Too bad the image management, tiling, and viewing software that Rumsey uses isn't open source, as far as I can tell.

The fine folks at NewsIsFree are offering treemap based visualizations of news headlines. The News Maps are restricted to a few categories, but I can see how per user News Maps could be useful.

The caveat is that this particular implementation needs quite a bit of usability work. The tool tip style of navigation nukes way too much of the map context. Something subtler and in place needs to happen. Following an article completely blows away all context you had, either by opening a new window or new tab. There's a lot of small fragments whose potential is hard to guage. A lot of the navigation is menu based where the menu winds up taking a lot of screen real estate.

This of course is based on my in-depth, extensive, one minute evaluation. There may indeed be other usage patterns that are more important.

All in all a noble effort though. Hopefully they'll get a large user population and improve the interface. The ranking algorithm they use takes into account some interesting data, including how many people click through on a story.

For whatever reason, Bloglines is really starting to eat into my FeedDemon attention. Part of it is that I have a Linux desktop at home and I use Bloglines as an alternative to all the Linux aggregators. None of them quite did it for me. The other part is that FeedDemon just isn't providing enough value add.

And besides, FeedDemon seems to be having severe problems with highlighting old entries as new. Everytime I come back to the Wired News feed, let's say, I'm told I've got 30 odd new items. Yeah, right.

Maybe there's a FeedDemon upgrade that fixes this.

Then again there may be no hope. One of the big reasons that Bloglines is really useful is that I'm mainly editing blog postings in a Web interface. Ergo, Bloglines gets a tab, the posting form page gets a tab, and it's easy to transfer links and clippings to a blog post.

Okay it's mildly lame to point to John Battelle to talk about Topix.Net but that's the only place I could find Rich Skrenta's summary of new features at Topix. In addition to Gary Price's review, Findory's Greg Linden has a few words to say.

The key nugget seems to be a much more sophisticated NewsRank algorithm. There are also some UI improvements, and I'm just amazed at the number of sources and feeds Topix provides.

Oh yeah, Rich Skrenta, Topix's mastermind, is a Northwestern alumnus. Well before my time though.

Update I probably already knew this, but there is a Topix blog with aforementioned details.

At least for my set of bookmarks the del.icio.us redesign is quite pleasing to the eye.

Greg Linden, founder of Findory News and Blogory, takes a closer look at MSN's recently launched Newsbot. Previously, I stated that Newsbot pretty much squashed Findory, but Linden's review indicates his product has more sophisticated algorithms under the hood. Ergo they still have decent prospects at carving out a nice niche. Mass acceptance is unlikely, but maybe MSN, Yahoo, or Google will swallow them somewhere down the road.

Previously, I've evinced some skepticism about Findory's chances, but I hope they succeed in the end. If they can make this stuff work I have to imagine the technology will find its way into webfeed aggregators.

Memorandum
From: Department of First Strikes
To: Larry & Sergey

We know you're still working on fixing e-mail, but first thing after the IPO, buy Bloglines.

Combine it with some of the smarts from Google News and make an aggregator that would get Scoble to switch. Laugh at Yahoo's My Netsca err... My Yahoo. Resuscitate USENET.

Cover costs with improved and increased Ad Words sales.

That is all.

Earlier this year the British Broadcast Company launched Get Writing, an educational site on creative writing. The site features space for aspiring writers to save their works, learning modules about writing, and forums for writing discussion.

They even have my favorite ruse for focusing user activity: contests!! But they also reminded me of something else quite useful that I should have remembered from teaching: easily generated challenges which students can readily evaluate themselves against.

I wonder if Philip Greenspun baked any similar ideas into photo.net

Memorandum
From: Department of Half Baked Ideas
To: Self

Tweak del.icio.us posting bookmarklet to ping the pingomatic. Have other services watch for our particular del.icou.us pings, and take action, e.g. archive current version of del.icio.us posts, or take a snapshot of the del.icio.us front, just for grins.

Mayhem ensues.

O'Reilly is set to launch a combo magazine/book (mook) called Make basically geared towards hacking stuff. Too bad the term hack has been so soiled by the popular press. I'm sure that name came up during product development.

It makes sense for a company that has a whole series of "hack" books. In the fast movnig world of consumer electronics and communication the standard book cycle is way too slow. This let's them cut down on the latency. I wouldn't be surprised to see every Hacks hack (cough, cough) wind up being a Make contributor.

If successful, Make could spark another burst of technological innovation. It's been a while since the home hobbyist was front and center in computing. The Web has killed most of the niche technical magazines, yet isn't quite right for transmitting instructional details to a wide audience. A copy of Make could inadvertently ignite the next Jobs and Wozniak to noodle off into their garage and cook up something insanely great.

I think I remarked to J. C. Herz once that "we're raising a generation of kids that will expect everything to be moddable."

The homegrown Readership Institute recently released a number of reports about the newspaper reading experience. Overall interesting stuff, but for online folks their recommendations for attracting younger readership through focused discussion and debate is particularly germane.

I was kicking this around with my colleague Rich Gordon for a bit and came to the conclusion that what's needed is a toolbox of "debate/discussion" widgets between the low bandwidth end of polls and the (too) high bandwidth open discussion forums. Sounds like grist for a class project.

Motorola is hitting the 300 Mbps mark with cellular technology.

Within 5 years, everyone will have a Parctab 100 times faster, and 10 times cheaper than Mark Weiser's that works everywhere in the world.

Of course, in analogy to Dorados and Altos, we'll have twice as crappy interfaces and be programming them in languages only half as good.

The empire finally struck back at Google News with Newsbot. Well at least made the previously known beta more widely available.

Microsoft is bringing their search experts in to do neato whizbang clustering and topic identification. As a bonus, they'll track what you read, throw it into the magic collaborative filtering cauldron, and generate better results, both universal and personalized.

Too bad they just smooshed Findory, although it should be pointed out, Findory probably has a different set of sources.