Mihai Parparita demonstrates how to integrate Bloglines and del.icio.us by extending Firefox to rewrite Bloglines' clip links. Now that puts me in heaven.
Serendipitously, I also found out about the greasemonkey Firefox extension, which let's you hijack the browser's internal representation of a page and modify it to your heart's content. Sort of gives you a toolbox to start doing the type of things the new Google toolbar does, except much more under your control.
I hate to be snarky, but somedays I really miss Suck.
I'm not sure it's possible to have a more buzzword compliant pile of marketing fu. Brandmasters? Brandmasters?!
A shining example of a false dichotomy. There's no reason personalization and socialization can't be blended. Two great tastes that might taste great together.
Rich Skrenta, NU alum and CEO of Topix.net has a good wide ranging post on "The Incremental Web", his term for the rapidly growing, frequntly updating portion of the Web. The gist? This part of the Web is the frontline for research, development, new services, and new businesses.
For the most part I agree with where Skrenta is headed. It's interesting how a range of approaches to interacting with the Incremental Web are emerging. We've got neanderthal RSS aggregators, hyper-personalized tools like Findory, topic oriented Topix.net, and who knows what else lurking in the hinterlands.
One point of departure I have though, is the intense emphasis on freshness. While it shows up in the "newspaper", one thing editorial staff is good for is reconciling and recontextualizing new information with past content. Think of all those retrospectives, backgrounders, recaps, and chronologies your local ink stained wretches put together. Something that usefully mixed The Incremental Web and The Reference Web would be mighty handy.
Jason Calcanis lays out a bit of his perspective on the backstory regarding the Bloglines sale. Besides laying out Mark Fletcher as playing both sides of the developing deal, Calcanis argues that the only lucrative market for aggregators is corporate tuned clients. In particular, sources that provide advertising supported content through feeds are not going to let aggregators add their own ads, at least not without a cut of the action. Not to mention that, another big content provider, CNet, has entered the fray with Newsburst
This portends the aggregator schism I've remarked upon earlier. Calcanis seems to make an assumption that aggregators won't severely mutate birthing a new market, an assertion which the folks at Findory might not agree with.
Calcanis post via Read/Write Web
Adrian Holovaty summarizes the new hotness at Lawrence.com. RSS and search alerts make it easier for folks to, gasp!, keep track of their news.
Some folks talk about citizen's journalism and shaking up newspapers, while others have done it.
Crow. Fork. Mouth. Insert.
Marry Hodder was absolutely right about Ask Jeeves buying Bloglines.
I'm starting to be persuaded by Gregory Linden's pitching of Findory as a next generation webfeed aggregator, with social intelligence baked in. I'm thinking there will eventually be a schism in how people use webfeeds somewhat similar to e-mail. There will be a teeming horde who interface relatively naively to syndicated content, maybe through a tool like Findory, ala Hotmail, Yahoo! Mail, Gmail, Outlook Express etc. etc. Then there will be a decent sized, but not huge, pool of folks who work with such content at a higher, more sophisticated level, and will need more powerful tools, e.g. NetNewsWire or Bloglines, analagous to hardcore Outlook, Eudora, and Oddpost users.
And just so it doesn't look like I'm smooching Findory's rear too much, I have to say I'm finding the personalized sports news feed practically useless. ESPN and ESPN Radio basically crush the Findory feed on timeliness, and all the news sources report on the same thing every day. Now if they had a sports oriented blog and news columnist personalized feed, that might work better for me. Give me more from the margins, not the center!!
Thought parkin:' An aggregator that allowed one to tag individual posts would also provide a nice mechanism for helping train Bayesian filters. This could help deal with spam problem in tag feeds. And we have ample evidence from e-mail that people can understand and learn this mechanism.
Mary Hodder is reporting that Ask Jeeves has bought Bloglines. I'm inclined to believe there's something going on, but the reporting of the event strikes me as odd all the way around.
First off, in regards to her sources, we only get "I learned it from a couple of folks.". Journalists can be accused of a lot of failings, but an ethic of documented sourcing and verification is a good property. "Even if your mother tells you, check it out". Also, there have been no independent verifications, which typically happens with professional news sources on "big" stories, and no denials from either Ask Jeeves or Bloglines. Finally, notice the nuanced language from John Battelle one of the best at this stuff: "claim", "much more to say on this later".
Besides, if true, given who I wanted to purchase Bloglines, I'm highly disappointed AskJeeves might close the deal. I've just never had a high opinion of their search and I'm suspicious of Bloglines becoming overrun by ads. I'm not even sure it makes AskJeeves that much more palatable. Who wants to buy a premier aggregator company with a second tier search engine attached?
From the given evidence, I'd say there was some kind of financial agreement made, but maybe not a complete purchase.
Good in-depth interview by Richard Koman with Stuart Butterfield, one of the masterminds behind Flickr.
Choice quote on APIs by Butterfield:
On the strictly practical side, I think we had one person inquire about using the SOAP version of the API. I don't know if any apps were actually built. There is at least one application built on XML-RPC. But all the others--I don't even know how many there are--are built on the REST API. It's just so easy to develop that way; I think it's foolish to do anything else.
The folks at Findory are doing lots of interesting things. Gregory Linden has a roundup of reactions to Findory Neighbors. The tool presents blogs related to a given one, by showing a weighted word list of source titles.
Findory Neighbors currently is a nice one shot whizzy hack. I'm probably too much of a wonk for it to be really effective, but transparency would be nice. I have some gut sense of why I get what's in my neighborhood and the weights, but some explanatory data would increase confidence and aid navigation. Transparency in terms of how the connections are determined would be nice, ala PubSub's LinkRanks explanation. Also, longitudinal information, e.g. monthly neighborhood snapshots, would be really interesting.
Okay, I must be working too hard. I missed that del.icio.us has added sparklines to the most popular listings. If you don't know what a sparkline is, term coined by Edward Tufte, the device is a small graphical data display that takes on word like properties. You start to recognize certain sparkline shapes as words that support discussion of data. Think of how a flatline on an EKG says "boy I was daid!"
Anyhoo, this is a welcome addition in determining why a popular link is actually popular.
Even though I don't use NetNewsWire regularly anymore, sorry Brent, I still like to pump it as the gold standard for desktop webfeed aggregators. Looks like it's approaching another milestone release.
Given some of the new features, integration with Bloglines and search by Findory, I will now engage in rampant speculation that a new webfeed and syndication powerhouse could emerge from the margins. A little consolidation here and there, you never know, combined with a little complacency from the big boys, and voila! C'mon, tell me you saw Google coming 10 years ago.
Okay, so don't bet the mortgage on that tip. But it's neat how some of the "smaller" players are working with each other.
Via Jeff Jarvis, then Rafat Ali, UK's The Guardian is offering a branded desktop RSS aggregator, called Newspoint. The tech company behind the app, Consenda is funded by Advance.Net and the LA Times. I've worked on a hyperlocal citizen's media project supported by the former, and the latter is a Tribune newspaper.
Very, very interesting. I wouldn't have thought for media companies to strike out into desktop app territory. While this scheme can support a better user experience, this is just out of their core competency. Maybe the gizmo is a thin desktop interface, with a mostly Web application infrastructure underneath. That's essentially what the Gush guys do, with Flash and Python. Besides, all of those guys have big Web and Web application knowledge in house, so why not do a Web based aggregator or outsource to Bloglines or Yahoo! for that matter?
Enough of the tech noodling though. This signals that some more of the big media companies are starting to "get" web syndication and feel it's territory worth experimenting in. They'd better before Yahoo! et. al. completely dominate the space. Also, I say some more because at the next to last Online News Association, here in Evanston, I alerted some media types that this stuff might be a big deal, and actually heard about a few interesting projects.
Final bon mot. If a local news company can make the aggregator branding hyper-locally oriented, both geographically, North Shore of Chicago say, and socially, e.g. high school sports fans, minivan moms, 20something partiers, they might have a winner. Throw it on a disk. Throw it on newcomers' doorsteps. Instant gateway to the community. The targeted news comes to you, courtesy of your friendly neighborhood newspaper monopoly. And advertisers would probably go bananas.
I'm as big a fan of tagging as anyone, but the rampage to determine if tagging can or can't generate useful emergent taxonomies befuddles me. At least in the del.icio.us incarnation, tagging is supposed to be highly personal. You're just marking your stuff. Who cares if it confuses the heck out of other folks.
Peter Merholz observes that social effects are overriding any metadata imperative. Tags are being used for signalling not taxonomizing. Instead of marking to help people categorize stuff, you mark to announce to the world that some piece of media is now available for a particular audience. Turning every tag into a post-and-poll stream like del.icio.us and Flickr, or tying into an engine with search feeds, like Technorati, makes it easy to track the signals.
The key difference is that with signaling, people only think in the moment. There's no sense of posterity, which probably leads to impoverishd metadata. I mean who wants to think about the right way to tag that party picture?!
Steven Berlin Johnson documents his desktop research tools, adds further depth on his blog, and then makes the penultimate connection, bringing together blogging as a research tool with desktop search. There's still a lot of mileage to go in this general direction, and it veers away from the bloggers vs journalist morass.
But if only I had a nickel for any time Vannevar Bush's Memex was invoked!!
Possibly the coolest looking IM client on the market, Gush 1.3 has gone final. Details of what's interesting in this release stashed elsewhere, though.
Through Lilia Efimova, I knew what Anjo Anjewierden was up to regarding visualizing blog content. When she was in town Lilia showed me Anjo's BlogTrace application, the subject of two papers. I've been hacking away at a similar concept, called a BlogScape, that's a little lighter on the analysis, although long term I'd like to incorporate many of Anjo's ideas. Here's to either a little friendly competition or enlightened collaboration.
One thing that's obvious, getting this to scale up to lots of blogs and webfeeds is going to take a lot of work.
Tim Porter, over at API Mediacenter's morph weblog, has six good ideas for blowing up the newsroom. The only faulty bit is the final question: Who among the traditional newsrooms is going to lead us into tomorrow by being the rule-breaker of today?.
Change will come from the margins and the outside, c.f. the comments on Porter's post. Look to the Lawrence.coms of the world.
They may not be "old" for all I know, but Patrick Logan and Bill Clementson have been hacking with Lisp for a long time. A lot of their weblog content has been great recently, if you're into programming language design and implementation. Also, they have a lot of knowledge on non-obvious mechanisms for Web programming.
Clementson conveniently packaged up his 2004 best blog postings. As an example, check out his thorough discussion of programming Web applications with continuations.
Logan has been fighting the good fight, making sure folks know there are a lot of good ideas in those obscure dynamic languages.
Bob Wyman is reporting that PubSub is watching 8 million blogs. About 4 million are active, according to him, although I couldn't quickly find a definition of active. Even more interesting is that the gap between number of blogs and overall output is increasing, according to this chart, and growth in output is relatively slow.
PubSub is a tool worthy of much more utilization and promotion in the blogosphere. Since it doesn't tie in neatly to publishing tools, the service, AFAIK, doesn't get nearly the press it deserves. But let's say you wanted to watch the entire blogosphere for links into BoingBoing, or heck, even your own blog, in real time. Where would you turn?
You might say Technorati, but they're only advertising 6 Mil blogs watched and they're trying to be a search engine on top of a watch engine. Plus polling RSS just isn't the right model for watching, even though PubSub does support such an interface.
PubSub might be the best kept secret for keeping an eye on the blogosphere.
One, two! One, two!
Is this thing still on?! Jeez, you'd think after two years I'd have learned my lesson and shut up by now. Time to re-blogscape myself.
To infinity and beyond!
Seb Pauquet launches a request into the lazyweb to put some reputation filtering on top of del.icio.us. The overall concept isn't too difficult to implement, the tricky bit is not hammering del.icio.us and getting your service blocked. As far as I can tell, to get the full effect (using all of a tags history) you'd have to scrape a bunch of del.icio.us pages.
This just highlights the maddening point that Joshua zealously guards the implementation and data of del.icio.us. As far as the UI goes, I enjoy his staunch position. Keep It Simple Stupid. As far as the data and APIs go, let a thousand flowers bloom. The nominal reason to deny access has been server load, prior miscarriages of usage, dealing with crawlers, stability, etc. etc.
I wonder if we're to the point that if you're going to build a web application, work on the Web services API real early if not first. Plan on getting hammered by people building bad, broken, wrong, and way cool applications on top of it. Your elegantly designed frontend is just gravy.
A little more thinking out loud. How big a group do you need for a del.icio.us clone to be useful? Someone's gonna come along and do an easy to install opensource knockoff, with an interesting API and standardized data import/export. Throw in an inter-delicious protocol and you'll have the Wiki story all over again.
It could happen!!
Provocative if nothing else, over at the monkey methods weblog Andrew Chen claims Technorati and Feedster are essentially dead ends as standalone concerns. The gist is neither enterprise is distinctive enough to avoid being swallowed by one of Google, Yahoo!, Amazon, or MSN. The most critical hit is that both engines have poor interfaces for actually seeing into that huge iceberg of content. If you claim to track the "news of the blogosphere" why channel people through a keyword search? In the comments, David Sifry (Technorati) and Scott Rafer (Feedster) obviously object, keying on "RSS is different" and their web services APIs.
I tend to agree with Chen. I honestly think that there are a couple of folks at each of the big boy's research labs who could, in a summer project, ride the house crawler and indexer to more interesting results on a large body of webfeeds. These side projects might not reach the scale of Feedster or Technorati, but once they internally figure a way to get something really distinctive out of webfeeds, then the mothership could swoop on the small fry just for the pre-existing database. Besides, I'll bet with Google and Yahoo engineers any day of the week. As Suw Charman pointed out, there's some nuts and bolts stuff that could use a bit of polishing.
Then again, maybe Technorati and Feedster have neato analyses internally, probably ad related, that they're keeping quiet. Besides I'm on record that Sergey and Larry should swallow Bloglines first.
Link parkin': Michael Jastremski is building Creative Commons based archive of stock photography entitled OpenPhoto.net. Bunch of neat ideas bearing further investigation in the about page.
Ben Hyde has been on fire recently. His latest musings regarding URL micro languages and rich, beyond keyword, tag query languages speaks to what the tagged up future might look like. I'm sure somewhere the DB community has explored much of the potential query language design space, especially in the context of SQL. However, a lightweight, user friendly, Web query language would be interesting. Maybe a reduced or tweaked XPath could fit the bill.
Also, I'm in the same camp as Hyde in that I really use this space more as a lab notebook then as a vanity press. As I tell anyone who'll listen, the data model of timestamped posts can be used for a number of different applications, which is one of the major conceptual wins of blogging. There's some structure, but the model isn't so rigid you can't bend it to other purposes.
Refining my earlier thought about tagging's secret sauce, tagging's popularity doesn't come from taxonomizing, it derives from signalling. By applying markers that can be monitored to media (URLs, photos) you can signal to other like minded folk the availability or applicability of that media, for a given community.
Technorati's tag search provides yet another means by which the signal is transmitted. Given Technorati's hig profile, tagging blog posts could become quite popular, but we'll need a better set of watch tools for exploiting the results. Here's hoping PubSub adds tag selection to their query language.
I may have spoken a bit too fast in my pessimism regarding a script/agent that periodically grabs my subs, picks out the ones "kept as new", marks them as read, and then does something interesting with them. Looking at the Bloglines API namespace it turns out you get a marker for items kept as new. HTTP traffic snooping with Firefox reveals it's a pretty simple request to toggle the mark.
I'm sure there are some caveats, but it seems like an end of the day info hoover a.k.a. knowledge aggregator could be built.
Happiness is getting your del.icio.us inbox back.
There's very little seriously rigorous blog research in my humble opinion. Susan Herring's brog, (we)blog research on genre, project was all over HICSS with good stuff. I have quibbles with some of their methodology, but at least it's clearly documented, the conclusions are argued from the data at hand and I could try to reproduce the results if I wanted to. Besides they have whizzy pictures!!
P.S. Tom Erickson and Susan Herring throw a pretty good mini-track at HICSS. I learned a lot and had a number of great interactions with folks like Jon Seely Brown, Dan Russell, and Jim Holland. Enough name dropping, eh!. If you can pony up the cash and time it's well worth the trip, independent of the location.
One issue I have with Bloglines is that I tend to mark items as new for further reading, commentary, or investigation. I could relatively easily post these to del.icio.us or put them in my clip blog, but I'm so lazy I don't want to even do that. Actually I blame it on the Bloglines clip interface which forces you through a popup dialog, a choice of where to stash the clip, and finally a confirmation window. Too much work!! Unfortunately, I have a big enough pile now of "marked as new" items that they actually make it hard for me to read with Bloglines. There's tons of old cruft intermingled with the new. Cleaning up is a daunting task.
I'd like to write a script that daily pulls down those tagged items, marks them as read, and in a blue sky world, fed them to a personal focused crawler. The crawler would do some clustering, context analysis, and automated search to build a compendium report for me. I'd pick it up the next day in my aggregator or stash it somewhere easily accessible. Call it knowledge aggregation.
This is only slightly more work than Findory, but maybe that little bit more semantic information (I really am interested in that post and here's some other feeds that I think are related to its source) might make a huge difference in finding relevant related stuff. And if search histories were programmatically accessible, hey more grist for the mill. At the very least the combo would help me separate my short term and longer term horizons.
The Bloglines API doesn't look very promising for making this happen. I'll keep digging, but combining post tagging, a web based webfeed aggregator, and a programmatic back end would be a big win.
Fight! Fight! Shirky vs Rosenfeld in a folksonomy vs controlled vocabulary smackdown. Ha, ha only serious.
One bit left out of the discussion is a minor, but crucial, implementation detail. The vanguard tools, del.icio.us and flickr, let folks watch tags through RSS feeds. Frankly, I think most folks could give a hoot about building an emergent taxonomy. They like being able to easily tag their content and they really like being able to watch how others tag theirs. People are still anxiously awaiting the return of the inbox on del.icio.us.
Forget the whizzy new vocab, I'm all about finding the whizzy new content.
Just stashing a whacky idea I had out here while looking at the Baseball Hall of Fame mechanism for determining who gets in. The process is this 1) every player is eligible a set time after their career, 2) they have to get a certain vote percentage (75%) to get in, 3) after a certain amount of tries they're eliminated, 4) if they don't make a certain threshold they're immediately eliminated and 5) voters have a fixed number of votes each round the number of which is relatively small compared to the size of the candidate set.
This process gives every player who ever played a chance, yet trims the pool quickly, and let's voters revise their decision, allowing for some thoughtful consideration. There's also an aspect of competition as you can stay eligible for a while, but you'll be competing against later, fitter newcomers as well.
Could this type of process be applied to media artifacts in a social bookmarking system like del.icio.us? This would be useful for identifying the "hall of fame" posts and highlighting them somewhere. You could also take them out of the consideration pool, so folks don't have to see them over and over. Newbies might get automatically introduced to them. On a per tag basis, this might be quite useful.
Tricky bits: what's a vote and how to deal with voter fraud? And could it be applied to other media artifacts. Or individual, personalized media streams. Hmmmm...
Just for Lilia Efimova since I know she'll catch this in her aggregator, I'm putting my HICSS Persistent Conversation Workshop presentation online. This was my response to Tom Erickson's request to think about how Slashdot might be redesigned to better support persistent conversation. The proposed changes are completely hypothetical and the whole presentation was mainly meant to spark conversation, not report results. Feedback appreciated, flames will be studiously ignored.
Merriam-Webster declared blog the word of 2004. Just one problem, they only have a noun form definition.
A colleague and I often joke about how "verbing weirds language", but the verb form of blog is what's really important. A new activity and mentality of personal, public, short form authoring has taken hold. Not just with words, but with photos, bookmarks, playlists, and listening lists. En masse, people are creating, ripping, mixing, burning, mashing, and blogging a new intertwingly mediasphere. Interesting times ahead.
With a few spare cycles, I enjoyed working through back entries of Bob Wyman's weblog. Wyman is the chief public face of PubSub.com. I particularly like how he tries to position PubSub.com as not a traditional search engine, although I find "prospective search" a bit tortured.
"Browse" is typically placed as the opposite of search, but I'm thinking "watch" might be more appropriate. And PubSub is one of the few, good, leading edge watch engines we have.
And just to be greedy I'd be all over any more technical details on how PubSub does that many matches.
Seb Paquet captures a potential emerging trend: commentlogging. If I get the gist, folks are using del.icio.us to record the places they leave comments across the blogosphere, squirreling them away for later examination. This also helps in monitoring the overall discussion.
By itself, this is an interesting usage of del.icio.us, but combined with Robin Millettee's durl service for monitoring del.icio.us URLs, or conceivably PubSub, some second order effects could kick in. I've been a big whiner about the impoverished nature of blog conversation versus classic USENET, but this seems to hint at further steps down the slippery slope of centralizing conversation. I'm pondering whether there's any significant improvement or extension but haven't seen any glaring wins. Still this might be an emergening phenomenon to keep an eye on.
Only problem is, making del.icio.us the sole conversation server, polled to boot, doesn't seem like a good idea to me.
I'm getting the feeling that webfeed aggregation is in for a big hype bubble this upcoming year. Vacuous buzzphrases like "the network is the blog" and "personalized information hypermarkets" are starting to get traction. Besides there's now a conference providing, "awareness, clarity, education, deal-making and strategic business opportunities surrounding the emergence of online media syndication".
Uh, yeah.
In the short run, webfeeds are creating more problems then they solve. First is the firehose effect. Frankly, current aggregators suck at ameliorating this problem. They're all a step backwards from Outlook or even a good USENET reader. Second, given that standard syndication formats make media more amenable to machine manipulation on the end user's behalf, advertising becomes severely threatened. Makes the monetization issue that much tougher. Not mention branding gets whacked with webfeed aggregators. Last but not least is the continued thrashing about regarding bandwidth issues. Currently aggregation can be seen as something of a screw for both publishers and readers.
Of course, during the upcoming chaos a few prospectors, and their suppliers, will get quite rich. However it will take a while for the real transformative value to emerge, sprinklings of the magic "social network" dust notwithstanding.
Darren Barefoot reviews Rojo, a relatively new Web based webfeed aggregator that tries to leverage social translucence techniques. I've got a one word response: ditto.
I find my Rojo account particularly uncompelling, and I get off on this stuff. The stampede of folks asking me for invitations, not, might also be indicative of overall interest from netizens.
Nugget 1: I was subscriber number two on Bloglines to the The Long Tail webfeed. That was yesterday.
As of this writing, there are 124 subscribers. Two orders of magnitude in less than two days. It's not what you know, it's who in the network knows you.
Nugget 2: One man's Long Tail is another man's white noise.