The web development and content management framework behind Lawrence Journal World's award winning sites is being open sourced as Django. Adrian Holovaty and Simon Willison were the primary developers. Python inside.
I was at the Chipy meeting Adrian wrote about, and can vouch that Django is pretty slick and intelligently designed.
Berkeley's getting another corporate funded, research lab. Marc Davis, of Garage Cinema fame, is starting up Yahoo! Research Berkeley, according to Bradley Horowitz. And he should know as Director of Technology Development at Y!. Here's betting you start to see some kickass, mobile and social media projects emerging from Yahoo! in the near future.
For those not into the academic research computing scene, there's a relatively recent trend where big computing companies create little knockoffs of Xerox PARC near computing research rich urban areas. Witness the Intel research Lablets, MSR in various places, and Google in Kirkland, WA. Heck, even locally Toyota has dropped a bunch of money on U of C to feed starving theorists. (I kid!)
As far as I can tell, none of them have had quite the synergistic energy and results the original PARC did, but dang if it ain't nice for the scholarly types who get hooked up with them. Depending on the arrangement, you get less teaching, more money, and access to a company's user base and infrastructure. Of course you're subject to the economic winds, you've got a bit more serious legal department, and there are fewer graduate serf...er...students.
Link parkin': PostgreSQL 8.0 has some easily installed, serious XML processing functions, courtesy of Tom Dyson and John Gray.
The American Association for Artificial Intelligence is holding a symposium on Computational Approaches to Analysing Weblogs (AAAI-CAAW) in March. Paper submissions are due October 7, 2005. Looks like a solid program committee and it's Palo Alto in March. Hmmmm....
Hat tip to Eszter's List.
I remember getting all excited about Robert X. Cringely talking about doing TV segments and distributing them on the Web. I thought I blogged it but apparently I didn't jot down my thoughts about how this could be a great experiment in P2P distribution (really!). Meanwhile, it mostly turned out to be vapor.
Well, it looks like Cringely's vision is coming true. You'll see a lot of links to the PBS press release, especially since the video is going to be Creative Commons licensed, but you can also get the dirt directly from Cringely.
Interesting post by PubSub's Bob Wyman on having to scrub certain links from MSN Spaces feeds. Those Microsoft RSS list extensions apparently wind up adding a lot of extra, but not timely links to RSS feeds. The arguably superfluous links, e.g. from blogrolls, will be excised from LinkRank calculations, but still forwarded on in PubSub search results.
This makes me wonder, aren't things like this better done with enclosures? This would have the benefit of keeping RSS relatively simple but also allowing for wider experimentation with other micro-formats. Having to worry about conflicting with various RSS factors has to be tough on extension implementors. Pushing the document out of band could relieve some pressure. Granted enclosures were intended for large media files, but there's no reason they couldn't be used for micro-format documents is there?
Taggregation application++. Shadows is yet another tagging application. Unlike, David Carpe, who pans the initial unveiling as agitprop, to me it looks like there's about 1/2 a new idea in there. Gathering all the information, (tags, comments, and ratings) about a particular URL into a Shadow page is a new twist on tagging apps, at least to me.
But Carpe's right, it's a pretty crowded dance floor and a I don't see a lot of reason users would pick Shadows to boogie with.
Robin Good has an in-depth review of BlogBridge, yet another entry in the crowded aggregator field. BlogBridge has some interesting features and philosophies, e.g. the hybrid Web/desktop strategy and focus on late adopting non-technical users. Obviously out of my demo, but I wish them good luck.
Meanwhile, in an extended interview with the developers of BlogBridge, Good gets to make some suggestions regarding the direction the product should go. He pushes on his idea of a Personal Media Aggregator, which has a lot of good ideas but suffers from a bit of feature creep IMHO. Also, a "walled self-contained communication structure" is a pipe dream these days, but maybe I'm reading this aspect incorrectly.
But as one who putters about thinking on aggregation, I'll have to dig into Good's missives a bit more.
Aside, BlogBridge is supposed to be GPLed, but how does that square with part of it being a Web hosted service?
Ian, please stop holding back. Tell us how you really feel about PHP.
There's an old Scheme-head saying, roughly paraphrased, "If you're thinking about using eval you're probably doing something wrong." First of all eval is hard to use correctly, hard to debug, and usually presents major security issues. Seems like a lot of eval abuse has seeped into PHP, amongst other sins.
Wow! This year's International Lisp Conference put on by the Association of Lisp Users featured a stellar lineup. (What!? No David Moon?). The speakers page includes some slides from the plenary sessions and the technical talks.
Biggest surprise of all? Curl Corporation is still alive.
Not that I particularly care about the success of the petition, but Jason Calacanis is campaigning for Google and/or Yahoo! to implement blog focused search. This raises for me an interesting question.
Is blog/webfeed search now a Google-class problem?
I use Google-class generically for "large, technology company with ginormous computing facilities, technical expertise and financial resources." You could replace the name Google with a small number of other companies e.g. Microsoft, Yahoo!, maybe IBM or AskJeeves. The point being there might only be a small number of existing companies that could "solve" the major existing, problems and issues.
Of course I'd be remiss if I didn't mention that evidence of blog search at Yahoo! has cropped up in the wild.
Update: More details on the Technorati slowdown, courtesy of BusinessWeek's Stephen Baker.
A thought exercise.
At one end of a spectrum we have the classic webfeed aggregator. You tell it where to look, and it brings back a bunch of stuff. Whatever it finds, you get. The aggregator never changes where it looks.
At the other end of the spectrum, you have Findory. You don't tell it anything, explicitly, and it brings back a bunch of stuff. Depending on the profile of you that's built, you get different stuff from different places.
What would an aggregator somewhere in the middle look like?
If you wanted to fool around with Web based visualization and need a live data stream from lots of sources, the WeatherBug API might be up your alley. This could also be a nice foundation for scripting programming assignments. Not hard to motivate the application and you don't have to explain the concept of weather to students.
Hat tip to Jeff Barr who also had a good thought on APIs and SPIs, service provider interfaces.
Mark Lentczner and John Horigan built ContextFree as an interactive environment for Chris Coyne's Context Free Design Grammar (CFDG) language. CFDG is a simple language for creating pretty pictures.
Just collecting these creatures to have an apiary of gentle slope, media oriented languages, ala processing.
Link parkin': blinklist.com, web based, tagged bookmarks app++.
Now that the frothing around Yahoo!'s MyWeb 2.0 has died down a bit, I can point to Greg Linden's roundup of pro and con voices as good one stop shopping to get a handle on what it means in the short term.
Despite my excursions into Drupal and WordPress, if you put a gun to my head and made me install a blog based, content management system I'd still go with Movable Type. Looks like there's a major MT revision forthcoming.
Jay Allen hints that there's over a 100 features and improvements. I'll be keeping an eye out for upgrades in a couple of areas: 1) groups and permissions, 2) finer grained control of entry ordering, sans extensions, 3) post metadata.
Still, MT to me has the best crafted underpinnings of all the blogging tools. The data model hits the sweet spot, and the Perl modules are finely honed. I haven't dug into the Php infrastructure, but I suspect it's just as nice, modulo my general misgivings regarding Php as a programming language.
Stumbled across Poor Artists Working For a Living, I don't know how, but it's a trove of computational art hacks.
A little stale, but in the NY Times, David Pogue gives Verizon's EV-DO a positive review for early adopters. I'm in the market for a new cell phone+service and can't decide if I want to go aggro-geek and get something fully kitted out with a top of the line data service, or stay cheap and use the bucks for something else.
Tempting, very tempting.
Link parkin': Michael Angeles' case study of integrating a social bookmarking service into a work environment.
As Evan Jones points out, the threadpool is a common tool for concurrent and parallel tasks. For Python, I've been using the solid, but long in the tooth, Xoltar toolkit. I'm putting Jones' workqueue module in my backpocket since it's of recent vintage.
L. M. Orchard braindumped his thinking on building a foundation for webfeed driven applications. The key things that come out are 1) feed crawling, 2) archiving data model + querying, and 3) client API. He focuses on the second point and similar to work I did with Jeff Cousens, an MS student, comes down on the side of a light mix of metadata in the traditional relational model, and stashing the raw data. As Jeff showed, this can scale to some pretty large feed collections on stock hardware.
Well, if you consider 1 TB disk stock hardware ;-/ which isnt' that farfetched. However, the key is to store as much raw data as possible. If you get your data model wrong you can always go back and reprocess it.
Nice article by Scott Scriven on using Python's syntactic overloading features to make RDBMS access simpler and cleaner. Of course this depends on your data model, but once you've got that straight most of the techniques Scriven presents are readily applicable. Highly useful for Web development in Python.
Linda Stone, former Microsoft social computing researcher, coined the term "continuous partial attention" and spoke at Supernova 2005. Nat Torkington took notes which have been widely linked. The gist is, as far as I can manage, that folks have electronically networked themselves to death, attention wise, fragmenting a scarce resource across too many connections. Next up, there will be a huge premium on committed attention through a small set of trusted connections.
As an aside, practically any work related meeting I attend, with more than three people, has an abundance of overequipped powerusers not paying focused attention. Faculty meetings, classes, academic conferences, you name it. I'm wondering when this reaches the point where committed engagement becomes a competitive advantage.
Stowe Boyd gets a little excited about Yahoo!'s MyWeb 2.0 and makes an interesting connection:
I continue to believe that the center of the social universe is the instant messaging buddy list metaphor: not just because I am biased toward real-time communication, but because human beings are the center of the socialized world.
The tie between social connections and buddy lists is great, the rest I think is a false dichotomy. I agree that tools for managing social boundaries, of which buddy lists are an important example, should be prominent in socially oriented apps. They're not the center of the world though, as the actual activity and media that connect people are what's really important. However, there's no reason a well designed app can't keep help users manage both aspects.
In most of these systems, there's "stuff" and "people". Both are important to end users.
But tools like buddy lists are sorely undervalued in many web applications. Witness the constant requests for private bookmarks in del.icio.us. Meanwhile, I admire Flickr's lightweight, but apparently effective, tools for managing social connections.
David Jennings does some comparing and contrasting of playlist sharing services. Not sure anything definitive comes out of it other than, Art of the Mix and WebJay are distinctly different. Still it's useful for reference purposes.
I've been trying to lay off slagging on their service, but the case against Technorati, as Jeneane Sessum points out, is building. I never quite understood the attraction since Technorati search results are typically unintelligible to me and the number of compelling apps that rely on Technorati's services is vanishingly small.
I'll have to see how Feedster, WayPath, Blogdigger, et. al. are doing these days, but despite obvious appearances there's probably still market opportunities in webfeed oriented information retrieval services.
Python For Fun is a collection of Python code for intermediate programmers, people who already have a handle on the basics of programming but aren't quite hacker grade or ace with Python. Chris Meyers' has put together a bundle of interesting starting points and kernels for exploration. There's components on plain old data structures, GUI programming, RDBMS manipulation, circuit logic, virtual machines and compilation. Best of all there are suggestions for further directions to take each module, essentialy providing a nice toolkit of potential programming assignments for the enterprising Python instructor.
As if developers didn't have enough catnip to occupy themselves with, Google Maps now as an API. As well Yahoo! Maps now has programmatic services.
And digging into yesterday's MyWeb announcement, Yahoo!'s also providing an API for that as well. Unfortunately, it doesn't look like there's any way to actually put data into a user's folders though.
Yowsa! Looks like Yahoo! is going! gangbusters! into social! software!, putting Google and MS (other than Wallop) on the defensive. Yahoo! 360 went non-invite recently and now there's MyWeb 2.0. The new MyWeb looks like a del.icio.us knockoff but leveraging and integrating existing Yahoo! features like group management and search.
And to my eyes Yahoo! is just about this far from putting together an innovative social content media aggregator that puts a hurt on the rest of the market. They've got all the pieces (Flickr, YTunes, MyYahoo, MyWeb, 360, a directory, and search), the smart dev talent, and the commercial motivation.
Hat tip to John Battelle.
Johnvey Hwang's del.icio.us director is a tour-de-force of web service and AJAX programming. The result is a cleanly designed interface for the del.icio.us power user.
However, this inudced the notion that AJAX just means Flash, except with stone age development tools, implemented slightly differently across the major browsers, and a retrograde user interface API? But look ma! No propietary plug-ins!
It's so broken it just might work.
Always collecting links to big piles of free media here at NMH. Our own federal government, at FirstGov.gov, has anted up, gathering links to lots of government run, public domain image sites. I mean where else could you find copyright free photos taken by W.E.B. Dubois while he was hanging out in Paris?!
I admit it. I was and always will be a Suck fanboy. I even bought the damn book! Those detestable forays into "really X invented Y" (e.g. Lou Reed really invented rapping)? Here comes one.
Suck really invented blogging. Not the technology per se but the attitude, many of the literary conventions, and the current reigning visual design ethos. Routinely updated? Yup. Oblique use of links as both citation and commentary? Yup. Snarky, personal point of view? Yup. Web friendly writing style? Yup. No subject out of bounds? Yup. 3 column layout, with thin content column? Yup. Unique link for each days content? Yup.
Suck wasn't a "blog" but it definitely was "blogging".
Don't believe me? Revisit some old classics like Dig We Must, News above the Title, and just about anything by Ann O' Tate.
For a recap in broadband, dive into Matt Sharkey's history of Suck, entitled "The Big Fish" which conveniently concides with the upcoming tenth anniversary of Suck's first post.
Apropos of nothing, I normally detest list making of the sort embodied by the American Film Institute's Top 100 Movie Quotes. Such lists have the dubious distinction of generating more heat than light, IMHO.
For whatever reason, this one caught my fancy, and it's hard to argue with most of the selections (another reason I don't like such lists). Moonstruck's "Snap out of it!" seems highly dubious to me though, along with LOTR's "My Precious". Sorry, the books did the heavy lifting on that one. Animal House got short shrift. I'd bump "Toga..." for "Fat, drunk, and stupid...", and "Was it over when the Germans bombed Pearl Harbor...". And from Apocalypse Now how can you leave out "Terminate... with extreme prejudice."?!
Pursuant to my commentary about navigating the long tail, Ryan Shaw of Berkeley's SIMS program pointed me towards a term, semionaut, coined by artist Nicolas Bourriaud. One of Bourriaud's claims to fame is the late 90's text, "Relational Aesthetics," where semionaut came from. Emerging from the art world, "Relational Aesthetics" is probably over my head, but I'm game to dig up a copy and see if I can get the full import of Bourriaud's commentary.
Interestingly, it's difficult to find a review or two of the book, despite what appears to be some fairly broad impact. An in-depth review in the London Review of Books, by Princeton professor Hal Foster, is behind a pay wall.
"Postproduction" also looks interesting.
Thanks Ryan!
Link parkin': I haven't had time to read it all the way through, but in his continuing series on how to actually build REST services, Joe Gregorio detours into generating sparklines RESTfully
No I am not 40, although I'm getting close. But apparently there are now over 40 Bloglines subscribers to this site. Yowsa!! And that's to the full content feed, so there may be a few others lurking out there, using the older summary feeds.
You keep reading, we'll keep posting. Actually, I'll keep posting anyway to augment my faulty memory, but thanks for the attention.
Billy Fowks Flickr Montager is one of the first applications of the Flickr API that actually constructs new media artifacts. I was initially impressed by the types of montages it created (there must be a relatively obvious algorithm somewhere for using images as pixels for other images), but even cooler is that the montage is actually a navigation device unto itself. Muy bueno!
MicroSoft had some "big" announcements about RSS at Gnomedex, which Dare Obasanjo covered in exquisite detail.
I put big in scarequotes because I can really only detect four things coming out of this:
Points 2 and 4 to arrive within an Internet age or so.
I might be worried about incompatible and propietary RSS extensions except RSS has such a simple data model and previous generators were often so crappy that consumers have been through this drill already. MS is just one more whizzer in the water.
Otherwise, yawn.
Don't know how I missed it, but Fredrik Lundh has been developing an extension framework for Python's built-in Tk based GUI toolkit. His module is named the Widget Construction Kit (WCK), and recently part of WCK's drawing mechanism has been implemented using aggdraw. In turn AGG is a high quality vector graphics library. Since Lundh is involved, WCK I'd bet the Python Imaging Library plays nicely as well.
The gist? Python is fast gaining an easily extensible, high quality, cross platform, interface toolkit. You don't just get a bunch of early 90's era widgets, but a framework for developing whizzy new interface elements. In Python no less!!