Scripting News

Today's a big shipping day

Sun, 20 Sep 2009 16:31:28 GMT

I didn't plan it this way, but a lot of stuff is ready all at once.

Over the day I'll link them in here. They will also appear in various other places such as the rssCloud mail list, on Twitter, and I'm not sure where else.

1. A rssCloud walkthrough for publishers. This doc is a shortcut for publishers so they don't have to wade through the michegas that cloud and aggregator developers do.

2. A self-documenting OPML reading list of real-time photo feeds. Lots of new stuff working here. I also added a photo feed for television, in anticipation of the Emmies tonight.

3. DNS for RSS feeds. I have a proof of concept app up, for you to try out and comment on.

N. And last but not least, and I had nothing to do with this one -- the Scobles shipped another Scobleizer last night. Baby and mommy are happy and healthy, and according to daddy, cute. Happy birthday Ryan! It's nice to have you on board for this crazy thing we call life.

Hooking the lizard brain up to the cerebral cortex

Mon, 21 Sep 2009 03:21:15 GMT

This is worth a special post as a followup to the earlier piece, it's so interesting.

I was talking with a friend today, he's an expert in DNS, and I said it's too bad people can't open their supercloud.org sub-domain in a browser. I understood why this wasn't possible:

1. The browser is looking for an A record or a CNAME, and we're setting the TXT record.

My friend who doesn't want me to say his name (and I respect that) said it might not have to be that way. We talked about connecting the DNS with HTTP and it was really intriguing. If you think about it, DNS is like the lizard brain of the Internet, and HTTP, while it is in some ways lower tech than DNS, is the cerebral cortex. To have them integrate is like bringing the Wright Brothers plane they flew at Kitty Hawk on an Apollo moon mission. But that's the way we do things in techland.

Then I thought -- wait a minute, what's to stop me from also registering an A record that points back to me, and then keeping a database locally that associates a name with an RSS URL. Then when you open your supercloud.org subdomain in a browser:

1. It comes to my machine.

2. It looks you up in the database I'm keeping that's a mirror of what's in the TXT records in DNS.

3. From there I get your RSS address.

4. To which I simply do a 302 redirect.

Hah! I had to try it. And guess what! The fcuker works!!

Try it out: http://davewiner.supercloud.org/.

And then go through the registration process, and try it out yourself.

This is actually important because everyone expected taht you'd be able to open the supercloud.org sub-domain in the browser. Now you can.

DNS for RSS feeds

Sun, 20 Sep 2009 17:12:44 GMT

Here is a Unix shell command that gets the address of my RSS feed: dig +short davewiner.supercloud.org TXT

It makes a DNS call to get the TXT record associated with davewiner.supercloud.org. That's different from an A record or a CNAME record. TXT records are used for things like this. That's why when you go there in your browser it doesn't go anywhere.

I have an app up for you all to try out, so you can have a supercloud.org domain for your RSS feed (it works equally well for Atom, or anything that can be parsed as XML).

http://dns.rsscloud.org/

Enter your Twitter username and password and the URL of your feed. It verifies that it is the correct password. (And it doesn't store it or use it for any other purpose.) And then if everything goes well, your username will map to a supercloud.org domain and point to your feed.

This is of course not rocket science.

But it does seem to work.

PS: Why this is interesting, possibly, is that it may give us a way to shorten URLs and make them more flexible if we want to build a loosely-coupled Twitter-like network with feeds distributed around the net. DNS is the little bit of centralization that the Internet, itself a loosely-coupled network, is built on.

Fuck you John Edwards

Sun, 20 Sep 2009 23:53:56 GMT

I read the front page story on John Edwards with delight, as I'm sure many others did.

Yes it is news, to those who ask, because he came close to being VP, and was a serious candidate for President and the press was ignoring the story until they couldn't ignore it any more.

I remember when Edwards came to speak at Gnomedex and everyone said he was great and he should have a webcam everywhere but in the bathroom and bedroom. They wanted the transparent candidate. He thought it was a good idea.

Now we know for sure, but I thought so for sure then, nothing about the guy was real. His story was good, too good. His wife has cancer. Meanwhile, according to the Times he's promising to marry the mother of his out-of-wedlock child as soon as the wife dies, in a rooftop ceremony with Dave Matthews providing the music.

This is the end of John Edwards, who if he could have kept the act up a bit longer might have been the Attorney General of the United States. He's fucked. And maybe we should reflect on just how fucked up we are that we go for that kind of nonsense. Edwards was our ideal of what a candidate should be. A commercial product advertised on TV with the judgement of a con artist. Just the kind of guy we want a heartbeat from the Presidency. Not.

Yes this is a front page story because who Edwards is says so much about who we are. Edwards is a good serious look in the mirror.

DNS progress, another wish for a miracle

Sat, 19 Sep 2009 17:20:47 GMT

There seems to be much progress on the DNS project I asked for a couple of days ago. Two or three developers seem to be approaching the point of deploying and the APIs probably are pretty close. I hope that when they surface we can try to get the APIs into agreement. The differences appear to be cosmetic.

There's another project I'd like to wish into existence.

I fell in love with the realtime API implemented by FriendFeed. Very simple, clever, reliable, functional, a kind of no-brainer that I wish more APIs were. The only problem was it only worked with FriendFeed, which is cool because there are a lot of interesting people there, but it got even more interesting when they released the back-end server, Tornado, as open source.

This is good for two reasons: 1. It's really good interesting technology, and 2. It doesn't come from Google, it comes from Facebook, one of their rivals. Diversify is always a good strategy in the financial markets and in the tech world too.

Anyway here's what I wish for. I'd like someone to show up in the rssCloud communithy with a Tornado server running that either is an instance of rssCloud (preferable) or is connected to an instance (i.e. receiving notifications). I'd then like to hook River2 up to this server through a persistent realtime connection so that it can be notified of updates through a NAT and firewall. Then of course I'd write up a Howto explaining how I did it, ask for feedback and hope that people create more software that interops.

How about it? Anyone up for making some history??

PS: I wasn't really a Deadhead, but I have lots of friends who are, and I admire much of the spririt of the community. One of the things they invented was this idea of needing a miracle. It comes from a John Perry Barlow/Bob Weir song. People talk about wanting "miracle tickets" -- that is, a ticket that gets you into something for free, like a Dead show. But the concept applies to airlines, baseball games, movies, private parties, you name it. It also applies to new Internet bootstraps. To get something like rssCloud booted, like Barlow and Weir, I need a miracle -- ever-ee day!

He has a million followers

Sat, 19 Sep 2009 01:02:19 GMT

Scott Simon is on the SUL and boasts in the first sentence of the description of his interview with Clay Shirky that he has nearly 1 million followers.

"Host Scott Simon has nearly 1 million followers on Twitter."

Scott Simon is an expert on Twitter because he has nearly 1 million followers.

He actually has 1,022,105 followers and follows 56.

I once replied to him on Twitter because he said something dorky about blogs on his NPR show. He said that people talk lovingly about their newspaper, it's "my paper," but no one says "my blog."

I told him that I have a blog and I'm proud of it, and I've never had a newspaper. I give money to NPR every year, but I wonder why. He's one of the smartest people there, and as you can see, he'd not that smart.

But I do love Radio Lab. They're on Twitter too. They don't have a million followers. Thanks!

The RSS channel-level image

Fri, 18 Sep 2009 16:45:07 GMT

A heads-up on something that's going to prove useful down the road, something you might want to start thinking about now. Credit for this observation goes to the brilliant Frenchman, JY Stervinou.

One of the things about Twitter that really works are the 48-by-48 images they call avatars. They quickly become symbols for the person. When someone changes their avatar it's surprisingly important. I changed mine from King Kong to Don Quixote and people started treating me better. Not kidding. People really want me to use my face, but it bothers me to look at my face all the time. When I figure out how to have two views of myself, one for me and one for everyone else... Anyway.

If we're going to bootstrap a Twitter-like network outside of Twitter we're going to need those avatars. And luckily there's a very nice place to put them, the RSS <image> element. It's as if when Netscape spec'd RSS 0.91 they knew that 10 years later we'd need this.

The only problem is that most RSS images aren't 48-by-48 (of course) and most of them aren't square. That's what you might start thinking about, creating a square graphic that looks good. Since many people and organizations are crafting Twitter versions of themselves, this should be a relatively easy thing to do.

BTW, I include the Twitter avatars in the cloud-enabled feeds I maintain for all the people I follow. Here's an illustration.

Radio Lab

Fri, 18 Sep 2009 14:27:57 GMT

Once in a while you come across a gem like the Radio Lab podcast. I listened in delight a few nights ago when an episode about death aired on KQED. An hour of philosophical and scientific stories about death, a subject we all must spend a fair amount of time thinking and dreaming about.

A couple of examples.

The three stages of death: 1. Your body dies. 2. It's buried or cremated. 3. Your name is spoken for the last time. They postulate that your soul is in limbo until you reach stage 3. For most people that day is the day you're joined in death by your last loved one. And this makes the point that fame, which so many peope seek, may not be such a great thing. After you reach stage 2, you have no influence on how your name is used. The poor farmer whose property was turned into a college after he died, and was named after him, now must wait until the college closes. And that might not happen for quite a while.

The next idea is about a dead language, Latin. What does it sound like? No one knows, because the last native Latin speaker died centuries ago. But think about pottery. It's spun on a wheel while wet. Maybe, just maybe while it's being spun, the grooves faintly record people speaking around the potter. So a Roman vase might be an ancient phonograph record and contain echoes of long-dead Latin-speakers.

These ideas aren't useful but they touch something inside me that I like.

Bravo!

Google, open communities, patents

Fri, 18 Sep 2009 12:34:32 GMT

When Google patents ideas that have been openly discussed and implemented in the RSS community, and then doesn't understand why this raises objections -- well, I'd say we have had a failure to communicate. At least.

I'm thinking about the patent that Google was granted on September 15 that covers reading lists for feeds. They say it covers other things, and that's probably true, and if so -- if they had to patent something they should have stuck to the new stuff. And I think there's a good argument that they should follow the conventions of the community and not patent their innovations, rather contribute them in the same fashion that others had contributed their good ideas.

It's as if Google ran Linux servers and used the fact in their marketing (no problem). Then five years later it turns out they had forked the Linux code base (which is permitted) and was marketing it under their own brand name (okay) and had not checked their improvements back into the community (there's the problem). This would be a violation of the norms of the community. True, it would also be a violation of the open source license, and perhaps we should have one in the RSS community. But in both cases these licenses would be hard or impossible to enforce. I don't believe the GPL has ever been tested in court. And in the case of open formats and protocols, who knows if such restrictions would even be legal. But clearly the "norms" part of the argument is stronger than the legal one. If Google wants to be part of the RSS community, it should be respectful of it. That means not using their size and legal resources to take what's good about our work, foreclose it, sell it as their own, and control others' use of it (which is the point of a patent).

All of this adds up to nothing if Google's lawyers are like lawyers everywhere, and they probably are. And if all their talk about being supportive of open source is just talk. But, on the chance that they're serious about wanting to work with and support open communities of developers, there are pragmatic reasons why they should be respectful and careful. And they have not been either.

This isn't all about Google...

The rest of us could have taken steps to prevent this problem. And we still can head off similar problems in the future.

1. There is an idea out their called peer-to-patent. The USPTO ran a pilot project that just ended. Seems like a good idea. Basically the patent applications are published before they are granted, giving experts a chance to comment on the novelty of the work, thus providing guidance to the examiner.

2. I've long argued that there must be a parallel patent system, a good one, that works more or less the same way as the USPTO's process with one important difference. At the end of the process the public owns the invention. The creator is given full credit for his or her work, which often is all they want. But a careful document is generated that creates a hole within which there will never be patents. Every one of these unpatents acts to combat the bad kind of patents. (This idea has already been widely discussed.)

3. And perhaps there should be cash awards for those unpatents, to create commercial incentives to produce novel ideas. That would go a long ways to counter the (imho invalid) argument that patents spawn innovation. I think quite the opposite. Most patents in my area are like Google's reading list patent. Filed after-the-fact by a big company, claiming the ideas of engineers working outside of large companies.

Google gets a patent on reading lists

Thu, 17 Sep 2009 13:08:13 GMT

Google will probably protest that the patent they filed on reading lists was defensive. But it's a bad patent, based on an "invention" that was already out there, being discussed openly on Scripting News at least a year before they filed it. If you can subscribe to a feed, why not subscribe to a collection of feeds? And when an item is removed, you no longer are subscribed to it, and when one is added, you are now subscribed to it.

Google should explain to the RSS community how they supposedly invented this and what their process was. If it turns out that we had prior art, they should tear up the patent and apologize from trying to hijack something that doesn't belong to them.

I worry about big companies doing awful things because they do them. It's dangerous to have Google control so much of the RSS infrastructure, from Feedburner to Google Reader to Blogger. They could easily pass information between those systems without sharing the information publicly. They could, right now, deliver features to users that competitors would be locked out of. All this built on ideas, formats, protocols and know-how that were contributed by others without any limits on how they could be used. And now this patent.

BTW, both River2 and its predecessor implement reading lists. Google Reader does not.

River2 in EC2

Thu, 17 Sep 2009 02:25:52 GMT

I'm working on an EC2 image that, when you boot it up, is a River2 installation. Since it's outside any firewall or NAT, it's ready to wire into the realtime feed network. Cloud-enabled feeds connect into Amazon's cloud. It's almost mathematical. I'm lovin it. I think we're getting close to the promised land. Hah.

I need a Domain Name Server with a REST interface

Thu, 17 Sep 2009 00:41:29 GMT

Here's what I need:

1. A server I can use to manage hosts for a domain that I own that am currently not using. I have many. I will pick one.

2. Ideally I don't even want to run the server myself. Someone from the community of people who read this blog who are interested in distributed realtime message systems and want to play a role in their development. This project will not use a lot of bandwidth or server resources. It's primarily for development. The other users will be geeks like you and me.

3. The server must have a REST interface. I need at least one call. It takes three parameters (that I can think of, there may need to be more). The three parameters are: name of sub-domain (something like george), record name (I'll explain below) and the value. The same call can be used to change the value. Probably should send a string that's a MD5 hash of all the parameters plus my password. Something like that. You can tell me what it should be, but nothing too fancy.

The record name is a DNS record name. Not A or MX maybe TXT. The value is the address of their cloud-enabled feed. So george.loose.ly would be the name of George Metesky's realtime feed. If you want to follow him, you wouldn't have to use his feed address you'd use george.loose.ly. The client would just do a DNS lookup to find his feed.

4. If no one is willing to operate the server, I'll operate it. It must be something that runs on EC2.

5. I need it soon. I want to start developing a prototype. Tomorrow? Friday? Please, no later than Monday. It seems like a fairly easy thing to do.

Anyway that's the idea. Comments welcome of course. And DNS gurus if I've made some egregious errors, please let me know, gently.

Do you have a cloud-enabled feed?

Thu, 17 Sep 2009 00:10:42 GMT

It's been over a week since Wordpress shipped their plug-in that added rssCloud capability. There are a bunch of feeds out there that are now cloud-enabled, actually a few million.

If you know of any especially interesting ones, news-oriented feeds that are frequently updated -- they could be pro or amateur, bloggers or BigPubs, commercial, academic or open source, left-wing or right, it doesn't matter -- what matters is that they are interesting and that they're real-time.

I'm looking for feeds to include in the default set of the next release of River2 which is coming together now. So if you know of some, either post a link as a comment here or send me an email with a link to the email address in the right margin on scripting.com.

What's wrong with this picture?

Wed, 16 Sep 2009 17:46:32 GMT

It took me a few minutes staring at it to figure it out.

With a hat-tip to David Rowland, iPhone developer, who sent it to a mail list I'm on.

Brad, let's get together

Wed, 16 Sep 2009 01:31:39 GMT

Brad Fitzpatrick did what I used to do, say what he really thinks in a blog post about RSS stuff. It's fine, but it is just his point of view. There are other points of view that are valid, like mine.

About "just happen to work at Google" -- come on, man -- how many people who don't happen to work at Google can add code to the following products: 1. Feedburner, 2. Google Reader, 3. Blogger.

Further, prior art is really important, it's how you keep the breadth of the pile of tech we create as small as possible, allowing us to build higher with the finite brain capacity each of us has. The cloud element was right there in the spec. And when we talked, you knew about it. So to say you never heard of it, well -- I think you had.

It's true -- I was pretty freaked when I saw the note at the top of your spec that RSS didn't matter. Sometimes I think Google really believes that. Now I'm here to say RSS does matter. You can't pretend it doesn't because it does. You blew every kind of smoke at it when we talked. That's really good motivation for a guy like me who takes pride in his work.

Now why did I get busy with rssCloud? Primarily because I wanted to remember how it worked. Once I got started, I remembered why I liked it, so I kept going. That's all.

Brad, we should get together and talk about bringing our projects together. This is what you were going to have to do whether I reactivated rssCloud or not, because RSS is there, and it's huge, and you were trying to ignore it.

Screen saver in Snow Leopard

Tue, 15 Sep 2009 17:20:50 GMT

Believer it or not there's been a lot of controversy about the screen saver in Apple's new operating system release.

I know this isn't something most people spend a lot of time thinking about, but I happen to think one of the nicest things about the Mac is its screen saver, cause I love high-res photography and one of the nicest ways to use great photos is to hang them on the wall on a 50-inch HD monitor and use the Mac screen saver to drive it. Try it sometime, you won't be disappointed.

That's why I was bummed when I did a complete fresh install on a Mac that's being turned into an art computer, first Leopard then Snow Leopard, when it appeared as if the "Choose Folder" option on the screen saver had disappeared. But I figured someone on Twitter would know what happened, and sure enough, Mike Murry pointed me to the new way of doing things.

There's a plus and minus at the bottom of the list. When you click the plus you can add a folder to the list. Nice little improvement. Used to be there could only be one folder, now you get as many as you like.

Just thought I'd leave a pointer here to anyone else who gets confused.

Monday morning stuffff

Mon, 14 Sep 2009 18:13:32 GMT

Chuck Barris used to announce The Gong Show as just some stuffff.

Of course our stuffff is very serious.

A new Rebooting The News with guest Dan Gillmor. One of our best. Dan drills into just what rssCloud is and realtime. He's one of the best interviewers out there. It was my turn this week to name inspirations, and I chose Young People, as exemplified by Blake Ross, Joe Hewitt, Matt Mullenweg and Joseph Scott. You can skip to the last five minutes of the podcast, it's worth listening to. Usually we choose older folk as inspiration, but we have to remember that youth, in the right hands, is itself inspiring.

I posted a proposed addition to the rssCloud walkthrough.

Typepad announced support for Pubsubhubub. I predict on Twitter that we will bridge it with rssCloud so support of one will get you compatibility with th'other. Earlier I admitted to being a dork and not seeing them as being in competition. After all, they're not commercial products. What I care about is decentralizing the realtime web, so we're not dependent on one company. Both methods accomplish that. The real problem is centralization.

Andy Oram at O'Reilly wrote a stirring ode to decentralization. At one time O'Reilly was a big proponent of P2P. Maybe they will be once again?

Zee is spelled Zed Eee Eee

Sun, 13 Sep 2009 17:40:48 GMT

My FriendFeed friend Zee is in town for a conference, and he's worried about how Americans will take to his name. At first I was puzzled. What's so hard about the name Zee, I wondered. I had only seen it spelled out -- Zee -- cause we've never talked verbally only digitally.

He explained: "I got a blank stare when I first said it in Starbucks. The barista asked, how do you spell that? I said 'Zed', 'E', 'E'. Received a blank stare. Then I said 'Zee', 'E', 'E', which then got him a little more confused."

Ahhh I get it now. His name is the same as the first letter in his name when you say it in American English. In British English there's no such confusion.

After a bit of back and forth I came up with a suggestion.

Make up some long incomprehensible name that begins with Z (how about Zarathustra). When you get the puzzled look, say "My friends call me Zee." They'll like that for two reasons: 1. They don't have to remember the name and 2. You said you want to be a friend. Americans generally like this.

Just a slice of life on the Internets.

Tornado

Fri, 11 Sep 2009 04:14:33 GMT

The former FriendFeed company now owned by Facebook did something very interesting today. They released Tornado which is the customized web server that runs the backend of FriendFeed.

I speculate in a thread on FF: "Just thinking out loud if there were a REST interface for the backend that worked like the REST interface for the client, I would be able to program both ends without having to learn the internals of your system. It would be really elegant, and probably wouldn't cost that much in overhead. I was able to create an interface to the client side of your realtime API in an hour or two. If I could sneak into the backend the same way that's all I'd need to at least put together a proof of concept. Does this make any sense?"

We need what their backend does to make the connection from rssCloud to desktops. This is something the FriendFeed guys mastered, and there's reason to believe it scales to the level we'd need since they are the guys who did GMail and Google Maps.

Interesting times we live in.

I also reminded people that when cool technologies are shipping everywhere it's not a time of death it's a time of life, as long as we have the web to connect our work, there's nothing exclusive about it. The engineers don't think we're wiping each other out, only the pundits and the hangers-on do.

Interesting times we live in.

And that's a good thing.

BTW, elsewhere on Facebook, our friend Blake Ross shipped Facebook Lite, which we heard was wonderful and are not surprised to find out is. Congrats all around!

Bad Hair Day at 7PM

Fri, 11 Sep 2009 01:00:55 GMT

We're having a special Bad Hair Day podcast which is a mini-reprise of last night's meetup in Berkeley.

Marshall is off tonight. I'll have two guests -- Doug Kaye of spokenword.org and Joseph Scott of Automattic.

Doug is working on a podcast aggregator that supports rssCloud.

Doug also founded IT Conversations that was a focal point in the podcasting bootstrap at the beginning of the decade. The Gillmor Gang got its start on Doug's network; we were inspired by his work.

Joseph developed the plugin for Wordpress that shipped at the beginning of the week. I haven't known Joseph a very long time, but we've already had a spectacular success, imho.

We'll talk about many of the things that were discussed at last night's meetup.

And for this podcast we'll have an IRC chatroom:

irc://irc.freenode.net/#badHair

Tune in at 7PM Pacific!