via Photochaining.

Tonight we had a very successful Web Montag Silicon Valley at Socialtext Headquarters in Palo Alto. About 35 people showed up, and listened raptly to cool demos and talks:

• mnemo, a cool interactive web search/explorer
• an ad-hoc demo of Jajah, free member-to-member phone calls
• Gernot Poetsch, talking about “James“, his multimedia aggregator/personal information butler (and why it will never run on Windows, only Mac!)
• Eugene Kim giving the background behind Hyperscope
• Mark “sIFR” Wubben, asking if Plazes could use RFID beacons as Plazes

All together, the talks made a nice portrait of the future of the web and hypertext. I had a flashback to our recent Wiki Wednesday, too, when Jeremy Ruston did his great TiddlyWiki demo — it fits right into the same next generation web presented tonight.

I started notes on SubEthaEdit and Mark Wubben joined in, and I’ve posted the notes to the Web Montag wiki: Web Montag 14.08.2006 Silicon Valley Notes. I apologize if I didn’t get your name right; if you were there, please edit the notes to improve them! :-)

The storyline: The 20th century, with its read-only labor, culture and politics, was weirdly totalitarian. The 21st century is a revival of a different (normal, given history) way to organize and produce within society. To fulfill this, we all need to practice free culture, and demand support of it.

He talked about a new thing, the problem that the various free culture licenses are not interoperable yet. (Interoperability – value diversity and opportunity over control.) This causes problems because it creates islands of free content that can’t be mixed together. So he suggests we set up a body — he recommends the Software Freedom Law Center — that can certify licenses as “close enough” and then have the various license providers add compatibility clauses to existing licenses so derivative works can be relicensed to a “close enough” license.

Out-takes, quotes:

• commodity layers invite competition
• PD-Wiki
• Written language, words, is the Latin of our times. The language of the people today is video.

Chris Messina is presenting. starts with some background about himself, and then how BarCamp got started.

First there was Foo Camp at O’Reilly’s campus in Sebastopol. Foo pioneered the ad hoc camp, but after a couple of iterations, not everyone could fit any more; O’Reilly’s space wasn’t infinite.

Tantek Çelik, Andy “Termie” Smith, Ryan King, Matt Mullenweg, Eris Stassi, and Chris wanted to have their own event, if they couldn’t make it to Foo Camp.

Tools: blogging; email & lists; IRC & IM; plazes; wiki

Plazes was where Ross and Andy bumped into each other, and Ross volunteered the Socialtext offices, so now they had a venue.

6 days

6 people

no experience

volunteers + $2500

= 300 people – $8.50/person!

= a really good time

Everything was documented on the wiki.

The rules of barcamp:

• talk about it
• blog about it
• if you want to present, write your topic and name in a presentation slot
• three word intros (
• as many presentations at one time as the facilities (and even surroundings) allow for
• no pre-scheduled presentations, no tourists, no PowerPoint
• presentations go on as long as they have to, or until they run into another slot
• if this is your first time at a BarCamp, you HAVE to present, or help someone else present

The Grid: a big, interactive sheet of paper with smaller pieces of paper taped on it

A culture of inclusivity – everyone participates, including remotely.

It spread from there — 30 or 35 barcamps in the last 11 months!

Future: BarCamp Earth on the 1 year anniversary — be there or be square.

I ask a question: how can you inject a little BarCampness into a staid, existing conference? Chris talks about BarCampSanFranciso, which was run right after Supernova. Kevin Werbach, the Supernova organizer, contacted Tantek and Chris to . Other examples: BarCamp Etech, OSCamp at Oscon, a pre-session at TED.

Chris says BarCamps end up being locally-focused and energize local communities.

Question: how well does BarCamp translate to other cultures (non-geek, for instance). Chris says it will translate well because it has already, and talks about a bunch of different kinds of camps that have happened (I didn’t type the list, darn). Event registration, Mollyguard, Wild Apricot.

Question: plazes? Chris gives a short demo.

We talk a little bit about wikis and camps. BarCamps feel a lot like a physical wiki; everyone gets involved in every part of the conference, including sessions, cleanup, etc. On the other hand, wikis have had trouble scaling, staying pertinent between events, etc. (there’s room for growth and innovation here).

BarCamp is a Community Mark. There’s no legal standing for this, but the community defends the mark. “Community response protects the mark.” “Sort of the Creative Commons of trademark.” There can be hacks of the mark, like BarCampTexas, which are okay if the community thinks they’re okay.

Question: how do you keep the local community energized after they get jazzed by a BarCamp? In 6 months, or a year? Chris and Tara talk about different local places where the community keeps it going, but it happens much more frequently, every week or every month, with dinners or other get-togethers.

Related: tequp, BrainJams, Coworking (getting it started so people can participate now, figuring out the way the expenses get paid, the “business model” if it could be called that for an open sort of org, later)

Jimmy talks about Wikipedia news from this year, and new Wikimedia stuff:

• article milestones, in hundreds of thousands of articles in each of a number of language
• the Seigenthaler incident – Jimmy deadpanned, “Apparently, there was an error in Wikipedia.” Of course, it was a nasty error, but it got fixed quickly; he talked about having to be on CNN to explain why Wikipedia might have errors, and how they’ll also get fixed. He shows a great graph comparing CNN and Wikipedia’s traffic at the time; CNN goes down, Wikpedia goes way up
• the good news of the Nature article comparing Wikipedia favorably with Britannica
• but also, he talks soberly about the background of the Nature comparison; Wikipedia got a little lucky in the comparison, because they picked science articles, not culture or something, and Wikipedia got started with geeks; they only evaluated errors, not writing style; and they picked articles that were roughly the same size as Britannica’s, instead of stubs. Wikipedia’s going to keep getting better, though
• Wikimedia Foundation is up to five employees
• Brad Patrick, general counsel and “interim CEO”
• Wikia funding, they were careful to pick investors, and let “a significant portion of our investment funds authorized to be used to support Wikimedia”
• Campaigns Wikia
• Wikipedia being the first element of the content repository for One Laptop Per Child
• board approving Wikiversity
• creating a formal Advisory Board, to interface with other institutions and the rest of the world
• Socialtext and Wikimedia working together to integrate Wikiwyg, our open-source what-you-see-is-what-you-get Wiki editing interface into Mediawiki. He told a sweet story of a friend who could have made great contributions to Wikipedia, but found she couldn’t because of the wikitext interface; he thinks Wikiwyg is going to be really important at helping more people contribute to Wikipedia. Right on!
• quality initiative – concentrating from growth and concentrate more on quality
• for instance, working on WP:BIO, better policies and taking a strong stand against unsourced claims, especially negative claims
• better image inclusion policies, especially concentrating on using freely-redistributable images
• “stable versions” – something that’s had a lot of thought, especially in German Wikpedia. simultaneously achieve two goals: let anyone edit things at any time, but also having a good stable article version for general public view. “One of the most important things we can do” to improve quality and the experience for general users.
• update on his 10 things from last year. I’m going to let other people blog each of these, but during the update of one of them, Jimmy said something funny, just the phrase “mission accomplished”. He said, “I guess I shouldn’t say that — it used to be a perfectly good English phrase, and now it’s ruined.” He paused, and then said that maybe it does fit: “Mission accomplished but there are still skirmishes every day.”

There’s time for one question: the policy of not having commercial information providers. I beg you to let vendors have one paragraph summaries about their products and services. Jimmy says this is sort of like the biography situation, and may take similar policies, and that this sort of decision is up to the editors and the community, of course.

Oh, time for other questions. Support for stable versions in the Mediawiki software itself? Jimmy says to talk to the tech folks like Brion and the German technical folks, but there’s just some simple support required to flag an article’s stable version. Jimmy emphasizes again that stable versions is something folks should really talk about here, because it’s really important, and suggests the technical folks do the simplest, quickest thing that would work, instead of trying to get a perfect solution right away.

The next question is about the ability for e.g., Wikiversity to promote the reputation of people who are good but wouldn’t usually get respect. Jimmy gives kind of a general answer, but compares free culture / wiki culture to the interstate system; it’ll change the rules and as a result, society.

A good keynote, lots of good updates, and Jimmy’s got a nice friendly, funny style.

“Why We Can Be Confident of Turing Test Capability Within a Quarter Century,” by Ray Kurzweil, excerpted from his book, “The Singularity is Near, When Humans Transcend Biology.”

Ray Kurzweil will present this paper at “The Dartmouth Artificial Intelligence Conference: The next 50 years” (AI@50) in Hanover, NH on Friday July 14, 2006.

The article is a richly annotated overview of where we are with AI now, and how Kurzweil thinks we’ll get to strong AI by 2029. Kurzweil fails to transmit his confidence in a particular date to me, but it’s still a very interesting article on hybrid human/machine problem-solving techniques, and a plausible path forward to strong AI.

Terminology note: the “G, N, R” revolution triplet Kurzweil refers to is “genetics, nanotechnology, robotics.”

There are longer lists of virelangues sur le web, but I got this one from Ginette Charette.

Christopher Alexander explains how a physical space is designed and then evolves to be useful and comfortable in his book The Timeless Way of Building (a highly recommended read, although personally I find some of the mysticism around the “quality without a name” a bit drawn out).

Some of the components are

• a collection of condensed architectural best practices (”patterns”) to be used in various situations
• articulating via “pattern language” a flexible plan tailored to a particular environment
• planning large structures in a hierarchy of scale: parts of rooms, rooms, single building, collections of buildings, cities
• explicitly “repairing” the design as people start to use the space in ways that were not foreseen
• creating without ego, to allow a “living” space to emerge for the people who live there

I think similar processes would work for online spaces, but there are some impediments:

• whereas architectural patterns of use have thousands of years of experience to draw on, online interactions have only been occurring for decades. It will take time to learn how people best interact online.
• it is only recently that a catalytic number of non-tech people have come to see computers and connectivity as a useful part of their everyday lives
• the “building materials” of online spaces at different scales — computers, programming toolsets, interoperation standards — are still rough, in flux, and often embroiled in market competition

We’ll get there, but it will take time. :)

Via Paul Barfoot.

It has problems with the new swarm download protocol BitTorrent, though. It regularly hangs, requiring a power-cycle reset to get going again. BitTorrent works fine, of course, but it’s a drag to have to reset the router a bunch of times during a big download.

Since the Linksys is running year-old firmware, and there is a newer firmware version 1.45.7 for it, I took the opportunity to upgrade before downloading the latest ISO of the stylin’ CD-based Linux Knoppix. Unfortunately, it didn’t really help; it still crashed periodically during the download.

The real problem, though, was that my POP3 mail checker stopped working after the upgrade. The POP3 checker runs on a Linux server — I’ll call it “Alpha”. It took some diagnosis, but here’s what I found.

• My connection to the Internet is over Comcast cable, with an RCA cable modem.
• Telnet on port 110 from Alpha to my mail server (also a Linux box) acted strangely. The remote server would connect, but not print the “+OK POP3″ herald until I sent a character. It behaved similarly as I issued other commands such as “USER” or “PASS”, with the general effect that its would be responding to the previous command each time. Needless to say, this confused the POP3 mail checker’s state machine, and was the reason it stopped working.
• I run VMware on Alpha, and instances of Windows within it. Oddly enough, from a VMware Windows 2000 instance — let’s call it “Beta” — running on Alpha, telnet on port 110 to the mail server worked fine, and didn’t run out of phase with commands.
• I did quick tests of HTTP (by hand) and SSH (using the ssh command) from Alpha. HTTP seemed to work okay (perhaps because the protocol doesn’t expect anything from the server to get started), but SSH didn’t want to start.
• To make sure something wasn’t set to a bad state after the BitTorrent download, I turned off all the computers and networking gear, then restarted everything. POP3 still didn’t work from Alpha.
• I downgraded the Linksys firmware to 1.44.2z, the previous working version I had been running. POP3 from Alpha worked fine again.
• To make sure it was the firmware, I upgraded to 1.45.7 again. The POP3 problem came back.
• Downgraded to 1.44.2z, the POP3 problem went away.

So, everything’s working okay again, except BitTorrent downloads. I’m still pretty happy with the Linksys box, as it’s served me well, but it looks like I should stick with 1.44.2.z for now. I’m going to copy this message to my blog, to the forums at BroadbandReports.com, where it seems like others have had similar problems with stalled email because of 1.45.7 (along with other problems; search for “1.45.7″), and to the Linksys comments page.