Movable Type 3.1 is coming out on August 31 allegedly. There are a lot of nice new features that make it even more useful, especially in news publishing environments. More plugin extensibility, subcategories, and scheduled posting are definitely a thumbs up. TypeKey and a limited user model still make fighting comment and trackback spam harder then they need to be.

MT 3.1 is also going to support dynamic rendering, allowing folks to trade off between fast, reliable, static pages, and whizzy, fresh, but slower pages.

Great! Except for the choice of Php as the language for implementing dynamic pages, at least from a systems design perspective. There's already a perfectly good scripting language, Perl, that MT relies on and has a beautiful API that the system exposes. Any of a number of Perl based dynamic page languages could have been added. I know Php is sweeping the nation, but I can't imagine this makes product management any easier. And besides, while I really try to be language agnostic, Php just seems poorly designed , through and through.

That being said, I completely understand the business logic. Tons of hosting services are Php enabled. The Web design crowd has taken up the banner. The language bears a surface simplicity that makes it seem more accessible than Perl. The engine is open source and actively maintained. It's basically a manadatory checklist item.

Heck, once you start down that route, why not just go to a pure Php solution?!

Steve Outing has a number of recommendations for building the ideal news web site over at Editor & Publisher. They mostly seem like old hat to me since we've tried a bunch out in the New Media capstone project. The key underlying theme though, get quality community contributed content, is still one near and dear to my heart.

Herewith, my 2 cents worth.

First, fill the search box with local shortcuts. Yahoo! has them for all sorts of stuff, why not add shortcut keywords for "coupons", "restaurants", "bears" (if you're in Chicago). Sure you have sections and menus to indicate focused navigation, but this would be another opportunity to add features in a place where you know a lot of people wind up. Also, your regular local folks will learn this interface and the site will be stickier for local information.

Second, focus user contributed content efforts within time limited contests that potentially use archival material. Have a weekly, "This Day In...", contest where folks can tell stories around a set of articles and photos from the archives. In college towns, have photo contests for homecoming. Win-win, all the way around, and if you can't sell sponsored advertising on top of this, shame on you!!

Need a caching HTTP server to put between the rapacious Web crawler you're running for the next great empirical blogosphere study? Think Apache and Squid are overkill for the task? Don't roll your own, use Julius Chroboczek's Polipo.

I'm just now kicking the tires but it's small, easy to manage, and seems to do what it says it does.

A few weeks ago, the proceedings of Into the Blogosphere started making the rounds. I refrained from commenting since I hadn't actually read any of the papers. I've managed to work through a couple and here's the skinny from an unabashed techno, engineering geek.

You have to be heavy into communication theory for the papers to be relevant and even so I'm not sure how "useful" they are. Mostly they seem to capture tightly focused snapshots of how people use blogs, or how people are thinking about blogs. But not of them struck me as seminal. Not saying they're bad, just that I think the ultimate impact of this collection will be quite muted.

I don't know what to make of this, but there are only about 40 comments, for 20 papers. There were 62 del.icio.us bookmarks as of this writing and three grand citations according to Waypath (Not sure I believe that). Any of you folks out there actually read the papers?

Truth be told I found the University of Minnesota's integration of Movable Type into the academic computing infrastructure the big discovery on my end.

David Carpe interviews Spurl's founder Hjalmar Gislason, revealing some of the philosophy that contrasts the Iceland based service with Furl.

Eric Zachte took Perl, Perl bindings for the Ploticus plotting library, devised his own special purpose plotting language, mixed it all up and out came EasyTimeline. The sample plots in the documentation are beautiful and apparently it's quite easy to add hyperlinks into the resulting images.

There are also Python bindings for Ploticus. A nifty programming languages class project would be to take one of the Python language toolkits and implement Zacthe's language or a different language design. One benefit is that you can readily see whether the dang translator/interpreter/compiler is doing the right thing.

So I signed up for a Furl account. Grabbed the "Furl It" link and threw it on my toolbar. Find an URL to furl and what happens?

I gotta rate the link, and think of a category to put in, plus maybe some comments, not to mention keywords, all in a popup.

Thankfully, there's a lighterweight bookmarklet, but man that's a lot of thinking to ask. Gotta say I like del.icio.us's clean and mean interface.Now I don't even hesitate to book mark stuff.

But I'm gonna give Furl some more time. I want to see how much of a win search is worth. Maybe I'll hack up a cron job to post my del.icio.us links to Furl.

Ran across the term stigmergy for the first time (sue me, I've been busy), in a John Robb post at Global Guerillas. Can't say as I'm really into The Global War on Terrorism, but the idea struck a chord in how to think about "communication" with social bookmark systems.

The term apparently comes from biology and the study of complexity generated by lots of simple creatures, ants and termites, using simple signals. Maybe this explains a certain fascination with del.icio.us as opposed to Spurl and Furl. The others are about information management while sharing and signalling seem to be secondary concerns. del.icio.us is all about those little signals and nothing else.

As a smaller world within the Web, social bookmarking strikes me as a sufficiently large, but not unwieldy, space in which to investigate some of these issues. Also, the arena is highly dynamic and a bit more amenable to automation then average Web crawling.

As to blog indices (blgodex, Technorati, Popdex) maybe their popularity (fading I might add) is a yearning for some kind of stigmergic totems.

Update: I don't claim any particular brilliance in attempting to make these connections. I'm sure a few have trod this path before me.

The Gray Lady joins the parade and provides publicly available RSS feeds. I'm sure somebody somewhere has done it, but it would be interesting to have a collection of newspaper/news organizations with RSS feeds.

Bloglines, a Web based aggregator, displays subscriber #'s along with a feed. I don't know if you can get those numbers in a programmatic fashion, but Richard MacManus raises some interesting questions about how they might be used.

Predictably it's the thought of yet another ranking/indexing scheme for weblogs. Not that I'm particularly against them, but maybe one day new types of metrics and measurements. Stuff that would truly get at measuring influence.

Here's one off the top of my head. Think of each item published by a source as an impulse into the media network. Try to monitor the spread of that impulse, either through direct linking and/or trackback analysis. Evaluate sequences of items to come up with a moving average of "impact". This combination of measurements can reflect being in "the right place" in the network, centrality to network analysts, despite having few subscribers. It also dampens one hit wonders.

Of course having lots of subscribers can help you be connected to, and maybe even in, those right network places. But low flyers and emerging stars could still get their just due.

Adrian Holovaty is a good guy. I even got the firm to fly him out here for a weekend.

Something tells me his Mozilla extension to fix All Music Guide is going to stir up a heap of trouble. It's your browser, but their content, so are you allowed to hack that last micron to your display screen? Lawyers are getting lathered up somewhere is my guess. Shades of Third Voice and it's ilk, the major change being that Mozilla/Firefox is much more easily (although not trivially) extended. If Mozilla ever retakes the world, look for a lot more hacks like this.

Mark Fletcher, the man behind Bloglines, has a blog: wingedpig.com.

I took a quick tour of the social bookmarks tools mentioned in Clay Shirky's piece on social link management. Mainly, I was looking for two things, RSS out in some fashion, and an API in.

Surprisingly, most of the services had at least an RSS feed for an individuals bookmarks, and often a number of other different outbound feeds, such as most recently posted links. APIs were harder to find. Spurl at least tells you they have one although they have to vet you before they'll let you in on it. The rest were remarkably silent.

Then with a flash of insight, I realized they all have some form of bookmarklet for posting to the service, and often for accessing other additional features like all the goofily named gobbledygook that Dude, Check This Out! gives you. With a little reverse engineering, therein lies some semblance of an API. Given that it's written in JavaScript, it ain't gonna be SOAP or XML-RPC.

One other thing I also learned, services like Furl and Spurl (are these two twins or what?) add archiving of bookmarked pages along with good old bookmarking. This is something del.icio.us doesn't support. I can somewhat see the utility, although for me, it's not often the case that a 404ed link is all that useful, even if its contents are stashed. Different strokes for different folks though.

Previously I stated that del.icio.us allowed you to subscribe to other users. Not only that you can subscibe to individual tags of a given user. Social connections can be made at a finer and more convenient granularity.

Urchin is a Web based RSS aggregator, implemented in Perl. There's plenty half baked takes on the same idea, but Urchin seems to have a nice clean logical design from top to bottom. Someone actually put some thought into how to build the backend of an RSS aggregator. In particular, it's DB schema was designed to cleanly represent feeds as RDF triples, and support the generation of RSS by querying the RDF store. Base modules have been speced to support ingesting any number of formats, while XSLT is used to generate output feeds.

Ben Hammersley gave Urchin a fairly positive review, although I would have liked to have heard if he actually put Urchin to use for an extended period of time.

At best Urchin could be the start of the aggregator platform I dream of. An extensible UI would have to be grafted onto it. At worst, folks could look at the DB schema for guidance on storage backends for RSS.

Waypath chips in with some blogosphere innovation, a pre-alpha technology called Topic Streams. Looks like it's in the same vein as a lot of the Topic Tracking and Detection work, sponsored by DARPA, within the information retrieval community.

So innovation and build out in weblog related search seems to be picking up, but where's the new stuff on the authoring end?

A theory. For broad adoption, a weblog authoring innovation has to be browser based. Browser technology is really stuck.

There is one possibility though, some enterprising soul really pushes the integration of wiki writing with blog posting. When I say push, someone does the hard work of making this happen within a popular blogging tool like WordPress or MovableType, and polishes it so non-techies can deploy and use. Every blog post becomes its own wiki page, and a non-HTML syntax becomes used for authoring. This syntax then becomes extended with some features that people really get hopped up about.

http://del.icio.us/inbox/user shows a user's subscriptions. For example, here's mine. It doesn't come in a nice structured format, so you might need to do a little screen scraping, but one could start building out an interesting network on top of this.

Note that in del.icio.us you can subscribe to both tags and users.

So of course, by beating the hell out of the service with a crawler, you could try and construct a global picture of del.icio.us. But more interesting to me would be how much social navigation is/could be generated using purely local connections. Other than the front page, there's not much of a shared global picture, by intentional design I suspect.

And just to add one meta level, you could monitor inboxes and generate RSS, so folks could subscribe to someone else's inbox and subscriptions. Basically, with a little effort, you can monitor a person's inflow and outflow.

Of course, Joshua Schachter has all the raw data at his fingertips too.

I'd like to say it was my special wizzy blogdance/posting about innovation that caused all this hot action, but I don't think anybody out there is buying it.

Feedster goes 2.0. Sounds tasty, modulo the worst permalink in the history of mankind.

Back in the old days, we used to call this feed splicing stuff, wait for it... aggregation!! Sure it's a bit smarter than your average NetNewsWire or FeedDemon, but it's the same general principal.

Seriously though, this is a really cool idea, merging "regular" blogging and photoblogging, and one step closer to my concept of a photomesh. Also, the concept of being an integration platform for differing "best" media blogging tools is a winner. The sweetest sounding part of the announcement, pending execution, is the development of APIs and standards.

Next up, I'm guessing, splicing social bookmarks. Flickr photos have tags. del.icio.us URLs have tags. Hmmm....

Who says there's no innovation in the blogosphere!!

Bloglines has a shiny new clip blogging feature, except it only clips items from your Bloglines feeds. del.icio.us works everywhere and seems like the right place to store all of this stuff.

Memo to self, turn clip blog into an RSS feed, then write a cron job to post new items into del.icio.us.

Clay Shirky surveys the current crop of Web bookmark tools, noting that what's old is new again.

del.icio.us comes out at the head of the class. One thing that would be interesting to know, from a geek perspective, is what API, if any, these tools expose. I'm betting on a burst of action on top of del.icio.us's REST API, but I'm just ignorant of what the other tools have.

Time to do some digging.

Well ya learn something new every day. I didn't know Python had a wizzy inspect module for poking around in objects, classes, stack traces, modules, etc. Who knew?

Python, is there nothing it can't do?

Is it me or is the "blogosphere", deadly dull, from a technology persepctive, these days? Name a new technical feature of blog tools/culture prototyped, announced, and widely adopted in the last year? I double dog dare yah!

Blog search engines? Feedster is sort of puttering along with nothing really exciting recently and reliability biting it in the butt. Technorati seems to be turning the corner on scale, speed, and reliability, but other than a couple of new indexes, anything to get the juices going? Related Technorati links? When's the last time you followed one of those? Besides, how useful is a search engine with a 7 day memory? You Waypath guys be quiet.

Speaking of indexes, anybody use Popdex, Blogdex, blogosphere.us for much of anything anymore? The HP iRank stuff has potential, but who besides HP, Google, MS, and Yahoo could implement it?

Photoblogs? Trackbacks? Old news.

TypeKey? Almost DOA.

SYOPML? Does it have legs? (Assuming it's actually still up)

Disabuse me of this notion by sending email to bmd at cs dot northwestern dot edu.

Poynter's E-Media Tidbits notes the launch of yet another youth oriented news spin-off, a combination of a print product and web site, that are different from the news organization's main paper and web site. The Lawrence Journal-World has been doing it to great success, and the last two New Media capstone projects here at Medill have followed a similar path. This is a trend to keep an eye on for a couple of reasons

First, these spin-offs look like they can actually be made attractive and profitable to a younger demographic. Even better they can serve as labs for news organizations to actually experiment with new Web techniques and technologies.

The spin-offs I've seen have been youth oriented (read entertainment), but a couple of segments that strike me as interesting would be a hyperlocal citizen's media product, and something targeted at seniors. A print adjunct to Silver Stringers should be a winner.

Part of the reason I fell behind on writing for New Media Hack, is that I've sort of fallen in love with del.icio.us, a social bookmarking tool that I've come to use more and more recently. As I've said before, del.icio.us isn't all that new as an application, but similar to my thoughts on Bloglines, the difference is in community scale. In addition, del.icio.us, also has a Zen like minimalism that carries over to its REST style API. del.icio.us's simplicity is limiting but freeing.

extisp.icio.us is the first hack I've seen built upon del.icio.us, but there's plenty more potential. For example, internally, del.icio.us could track link copies between bookmarks. The application also supports subscribing to other users. Collaborative filtering anyone?

Also, here's one of my patented half-baked ideas. Provide a calendar based interface on top of a person's del.icio.us bookmarks. Within each day, week, month, provide a global analysis of the URLs that came across the transom. Add some local analysis of the URLs an individual bookmarked to generate a tiny bit of social navigation. For bonus points, implement on top of a jazzy, infoviz toolkit like prefuse to make it sexy.

Anyhoo, there's probably a paper or two in there for some enterprising academic.

If you'd like to read New Media Hack right in your comfy old aggregator, there's now a full content feed for you.

If you're into social software, webfeeds, aggregation and so on, you really should be keeping an eye on Bloglines, the premier Web based feed aggregator. They recently redesigned and made an effective Web interface even more usable. Just rolled the dang thing out, and everybody benefits.

Bloglines also added "clip blogging". You see an item in a feed you like, you click a button, it goes into a blog. Big deal you say, plenty of of aggregators can do something similar. Yeah, but they're not holding the data for tens of thousands of people. And since Bloglines owns the interface, if clip blogging takes off, they can start adding fun stuff like group clip blogs. I worked on a paper submission last year (rejected, wah) which posited on some interesting UI stuff that could be done in such an environment. The problem is getting enough people together to run a decent experiment. Well there's at least one place it could be done.

Bloglines: the Google of aggregation?

Philip van Allen's piece on interaction design for online news was sitting around in a Firefox tab for quite a while. I was just about to nuke it but instead dug in a bit. Good choice.

van Allen forwards a notion of "productive interaction", where interacting with media is not so much consumption as production. This has huge potential online where the ability to interact is so much greater than with paper.

What's being produced? There could be other media objects, or personalized configurations of a media space, or, at the highest level "sense". Through interaction, the user produces an understanding of the media, i.e. they "make sense" of the work.

Productive interaction as a design philosophy is proposed by van Allen as a way to get online news out of its current local maximum. News sites are starting to fossilize. However, productive interaction is somewhat perilous in that users can actually fail to produce what the author intended and it takes a lot of up front thinking about what the user can do. This will probably have to be explored out on the fringes of the media commons before the big boys even think of adopting it.

I'm not sure if there's really anything new here, hypermedia folks have been struggling with similar issues for years, but put in the context of news a little bulb went off in my head.

All the popular aggregators are starting to look the same, and don't do much beyond supporting feed reading. I've probably said this somewhere before, but it would be nice to have an open, domain specific, scriptable (internally and externally) platform for slurping, slicing, and showing webfeeds and equally important meta information and analyses of these feeds. Much in the way that Emacs is not only a text editor, but a platform for building applications that do a lot of text munging.

Memo to self: find more round tuits.

About a month ago, OJR ran a feature on RSS in the news industry. The central concern seemed to be that if people used RSS to read the newspaper headlines, they wouldn't visit the news site home page, where all the ads are.

I don't know why, but this engendered the following stupid media trick. Differentiate your dang RSS feeds!! Think of RSS subscribers as a completely different, self-identifying audience for your content. Point towards different stuff. Add items with special offers, a.k.a. targeted coupons. Grab interesting stuff out of the archives and make it free for a time. Turn your feeds into a premium service for those who want more than the front page.

Just a suggestion.

Okay, so I fudged a bit. I'm back to life for real. No more excuses. What's a month's worth of posting between friends?

Looks like I managed to recover this blog, among other things, completely intact.

Didn't even have to break any URLs. All praise Murphy, Apache, Linux, FreeBSD, and MySQL.

In a stopgap version of this site I mentioned how the end of the quarter, two presentations, a paper submission, and costarica getting hacked put the kibosh on posting around here. Expect things to pick up a bit.

Hmmm, I detect an emerging meme: categorization is dead. Google advertises that in gmail you don't have to organize your mail. The guest lecturer in my class on Monday, remarks on how categorization is too hard for users. And Philipp Lenssen documents a few other ways categorization breaks down, and by association navigation in general.

Zinc Is Not Canvas, it's a hell of a lot better than the default Tk Canvas.

Yee, Fisher, Dhamija, and Hearst, a combination of Berkeley CS and SIMS folks, devised some effective animation techniques for radial tree layout, a few years ago. Radial tree layout is a focus+context visualization technique for showing graphs. The only place I've seen their methods show up is in the prefuse toolkit, but I'm not really deep into infoviz. Mainly parking so I can dig into the roots of radial tree layout.

One of the nice things about Medill's masters program is that one quarter is typically complete immersion into a project. 6 new media students launched goskokie.com on April 23rd. They've been adding features, generating content, pounding the pavement, and thinking up business models like mad. For less then a month's worth of time, those ink stained wretches have put together a damn good site.

While I've been doing occasional low tech troubleshooting on some plumbing, make no mistake they've done 95% of the hard work. One student could be considered a power user, but there's no hacker hiding behind the curtains. And this is a real live site, not a screenshot special.

Their most judicious decision? Selecting geeklog as a platform. Nominally a weblog tool, the dang thing has easily installed plugins for just about everything you can think of.

Link parkin': Tamara Munzner's archive of her papers regarding layout for large scale graphs, espceially her thesis on hyperbolic layout.

Hmmmm, could Bloglines become the aggregation platform I've been dreaming of? With a proper Web services API, interesting frontends can be experimented, including embedding within Web browserss, such as Mozilla. It's only simple notification, but Chad Everett's Mozilla Toolkit for Bloglines is a start. I may have to sign up with Bloglines just to kick the tires.

Link parkin: Six Apart's press release regarding Movable Type 3.0. Seems like there's a bit of a hubbub regarding licensing fees, but what I want to know is whether they improved the rather impoverished user model. MT 2.x didn't even support a notion of groups, much less assignable capabilities. I don't quite understand how you can go corporate without such mechanisms. Large groups and large #'s of blogs pretty much demand them.

Google is revamping Groups/USENET, with Atom feeds, amongst other features. Yes, syndication has come full circle and reinvented USENET, ... with HTML, ... and worse newsreaders.

Okay, having a kickass search engine for USENET is a win, but still, I miss my GNUS, sniff. (Northwestern IT whacked it's USENET server this year, losers).