February 2008 Archives

I needed to see the Natalie Portman SNL rap, so I went over to YouTube and plugged that phrase into the search box. the first hit was pretty surprising:

 

Contra Costa County BART stations to get library book vending machines.

Every now and then somebody comes up with an idea like this that seems obvious in retrospect. Hopefully they will carry a wider selection than Danielle Steele and The Da Vinci Code. To whom do we talk to get comics in there?

I helped fund this project at an elementary school near Charleston, SC through DonorsChoose.org. A 4th grade literacy teacher needed to put together a class library containing popular modern children's titles. A few days ago I got a package in the mail with a note from the teacher and some cards from the kids.



Totally unexpected, and extremely awesome. Apparently some of the kids were nearly fighting over the Harry Potter books :D

Full set here.

ps: I'm so awful at dealing with photos that I can't even make iPhoto work right!

I got a bug up my ass this afternoon to figure out why Twitter was giving me a 406 response when fetching my updates feed. To recap, I could get the feed successfully from the browser but not from the command line.

Inspecting the request headers sent by the browser (thanks to Penny for turning me onto the Live HTTP Headers Firefox addon), I noticed that the browser was sending a shitload of cookies. Light bulb time. The browser had an authenticated session but curl did not.

After a bit of digging, I found this Twitter account setting: "Protect my updates - Only let people whom I approve follow my updates. If this is checked, you WILL NOT be on the public timeline." Unchecking this box allowed curl to receive a 200 response to subsequent feed requests.

My question to the Twitter people: why not a 403 response? 406 indicates that none of the media types indicated in the request's Accept header are acceptable to the server for use in representing the resource in the response. 406 has nothing to do with access control.

Anyhoo, my tweets should now be showing up in the action streams on the front page.

Mine was the spicy chicken sandwich. Mine buddy's was the red beans 'n' rice. The venue was Kate's Kitchen in the lower Haight.

Happy Valentine’s Day to all. And to those who hate the day, I say this: Valentine’s Day is a Christian corruption of a pagan festival involving werewolves, blood and fucking. So wish people a happy Horny Werewolf Day and see what happens.

http://www.warrenellis.com/?p=5608

At least one friend is grumpy with our esteemed senator...

Apparently Dark Horse is now doing Dark Horse Presents over at MySpace. Lori pointed me to this Joss Whedon/Fábio Moon story, which is actually extremely fucking funny. Very cool that DH are doing exclusive online content, but too bad it's on MySpace - there's no feed or any other mechanism for notifying me when new issues are posted. MySpace is a technology ghetto. Comic book publishers are always behind the curve on this stuff. They need to get with the times and hire people who have some idea of what's happening in the industry today, not two years ago.

I just realized that I never linked the "Chopper Read" HTFU video. Thanks to Penny for this gem.

 

What is up with Twitter lately? First it was giving odd responses to feed requests. Now its Jabber/GTalk bots seem to be offline. I got all hot and bothered when Penny clued me into the fact that I could funnel Twitter through BitlBee into Irssi, but since I can't get at Twitter with Jabber, I guess I'll have to keep ignoring it, since I don't need peoples' "I'm on the shitter" "status updates" sent to my phone.

Tonight I ran across this plugin that enables Movable Type to act as an OpenID server. I had been using phpMyID as recommended by Sam Ruby, but I never liked having a login process with separate credentials just for my OpenID. With this plugin installed, whenever I log into a web site with my OpenID, I get redirected back to my MT installation to log into it. I reuse the MT login system and credentials, and that little bit of PHP code is gone from my site. One less thing to have to track, upgrade and potentially report bugs on.

I'm starting a new job on Wednesday at Metaweb. I'll be working on the data team building tools and applications that help make the data in Freebase better. This is a new problem space for me, which is one reason I'm excited; another is the sheer amount of ambition that these people have. They are talking about organizing and structuring all of the world's knowledge. I can't even begin to imagine what that really means, but they seem to have convinced some rich guys they can do it. So now I'll have to learn Python, and I'll have to work in an office again, but I guess I can use these changes of pace and scene. After working mostly at home for three years, I was starting to have conversations with the walls.

Today I updated to Movable Type Open Source 4.1 with the Action Streams plugin. The most interesting result is an additional column on the main index page with a list of crap I've done recently on other web sites. I haven't done a lot recently anywhere other than del.icio.us, but as I start using some of these services more, you'll see entries show up here. If you are so inclined, you can click the "More..." link to get to an archive page with a link to an Atom feed of the action stream.

There's also a list in the far right column of the same page with links to my profiles in assorted locations around the web. The profile links bear rel=me, so once Google re-indexes the page, I should be able to use the Social Graph API example apps to see these connections. These links are automatically maintained by the Action Streams plugin and replace the static list of similar links that I maintained by hand.

One thing I do quite often is read new books. Action Streams supports Goodreads but not LibraryThing, so I made the switch and so far am pretty impressed. I also listen to a lot of music, but Action Streams doesn't seem to grok Last.fm's recent tracks feed, presumably because of the volume of entries that it would generate. I can live with that.

Update: There does appear to be a bug with the Twitter support though. I'm getting system log entries that say

Could not fetch http://twitter.com/statuses/user_timeline/bcm .atom: 406 Not Acceptable

Not sure why that space is in the URL (I deleted and recreated the profile to make sure there was no space in the username), but it doesn't seem to matter; I get the same response without it. Anyway, if I plug that URL into my browser's location bar, I get an Atom feed in response as expected, but if i use curl to request that URL, I get the 406 response above, regardless of whether or not I include an Accept header or what value I provide for it.

Finally, something useful for looking at social graph information in public data: Google's Social Graph API. I've talked  to a lot of people about the potential for encoding structured data in the public web and the interesting tools that are starting to crop up for finding that data and letting people act on it. Microformatted events and locations are cool and all, but the relationships between people are just as interesting, and not just as vectors for empty viral marketing bullshit advertising stories.

I want to be able to stalk my friends without having to lock myself and everybody I know into using Facebook. By "stalk", I mean things like: finding out which rock shows they're going to this month, what books they're reading, what trips they have planned, and the experiences they've recorded with photographs. And I want to get this information straight from the web sites where my friends are publishing this information, not from some intermediary aggregator who forces me to use their web site to see the info, jams ads in wherever it can, and asserts some sort of shady ownership over publicly available information. Nor do I want to have to sell my soul to some centralized profile management system that everybody else has to agree to use. I want full control over how I express my relationships to people, and I want the same for everyone else.

Hopefully this announcement will stimulate the makers of web publishing software to build in support for XFN and FOAF. Digging this info out of the public web is only half the problem; getting the info into the web in the first place is the other half. As an example, Movable Type should include a module or widget or whatever their jargon is that lets me specify links to my public profiles and automatically insert the appropriate XFN magic; same with FOAF and links to my friends' profiles and sites. Hand coding this stuff is tedious even when it's as simple as rel="me", and the publishing interface should hide the details from us, since we're not all geeks, but we all want to blog. (MT should also have an authoring widget for dropping a microformatted event into a post!)

Tim O'Reilly has more in-depth comments at O'Reilly Radar. As he points out, beyond this great first step, we need ways to get access to the information encoded within each profile pointed to by one of these relationships, and ways to follow the same relationships into and within private data sources. I'm not smart enough to speculate on what these tools will look like, but I'm more hopeful than ever that they are on their way.