Scripting News

Berkeley is a small town

Thu, 12 Mar 2009 03:36:07 GMT

Tom Hunt a Berkeleyite since the 60s, who makes a living helping people keep their computers running, says Berkeley is a small town. If you look at the University as a factory, everything else is just a little place where everyone knows someone who knows everyone. I've now lived here 2.5 years and it's not unusual when I go out to meet two people I know. It's a small town with lots of creative people who like to use their minds. My kind of place.

Anyway -- there was a Laconica hackfest here a couple of Saturdays ago, at the open working space on Shattuck near Ashby. I had never been there, but I know a bunch of people who work there. Turns out a friend, Ken Sedgwick, was hosting the meetup even though he doesn't use Twitter or Identi.ca very much. But Ken knows how to package up Unix software so that it's easy to install and maintain. He's a consultant, that's one of the things he does for his clients.

So I went to the hackfest with a mission -- I wanted to create a Twitter that anyone one could install for themselves and host in Amazon's cloud. I have a theory that Twitter can be like Lotus Notes, a workgroup application that installs like shrinkwrap software did in the 80s. I want to learn about this, and hopefully -- if the theory is right, help start a billion Twitters to go with the billion weblogs we've got running now. Or something like that.

We had a dinner last night at China Village on Solano, eight people showed up, Ken told us where he was at. It sounded great, but he kept saying how much more there was to do. Even so he let us try out what he had, and it was a lot easier than most of the tech I have to crunch my head through (I was thinking of OAuth, which took two weeks). I was able to get my Twitter running in 10 minutes. Nothing to it! (Now, I have a little experience with EC2 and that probably helped. But it's still really easy. A tech type could get through it in no more than an hour.)

So here it is...

http://home.smallpicture.com/

This is going to be the Scripting News community Twitter if such a thing is possible. Create an account if you want. I have no idea if this is permanent or if we'll have to start over at some point. But I'm proud of the work that Ken did, and excited about the possibilities for the future.

I don't want to point to Ken's HowTo until he says it's okay. I think it is. But you never know, we'll do things in the right order. But know that Berkeley is humming, we're creating some good stuff. Glad I moved here.

An alternate OSCON?

Wed, 11 Mar 2009 20:58:32 GMT

OSCON == Open Source Convention.

Pretty sure I was at the first one, in Monterrey. It was described to me by Tim O'Reilly, who puts it on, as a place where all the open source platforms could come together under one tent. Perl, Python, PHP, Apache and many others.

It's been in Portland for the last few years, and this year it's moving to San Jose.

I had hoped to lead a discussion at this year's OSCON about porting Frontier to Linux. Frontier is the runtime environment that the OPML Editor builds on. It's an object database, scripting language, outline-based editor and database browser, debugger, multi-threaded runtime, verb set, Web CMS, TCP stack, built-in web server. It was the environment that XML-RPC, SOAP, RSS and of course OPML were developed in. All this and the whole download is about 5MB and it installs in a minute. It also has an RSS-based updating mechanism and most updates are "hot" -- meaning you don't even have to relaunch. I love this enviroment, I built it starting in 1988 as the last programming environment I'd use, the one that had everything I wanted, and that's what it is. And because the early development was done so (ahem) early, it was designed to run well on 1Mhz machines with 1MB of memory. As a result it fits really nicely in today's machines.

Now, it runs on Mac and Windows, but I really want it to run on Linux -- so I proposed a session at OSCON to discuss this and see if I couldn't recruit people to work on this. Unfortunately, yesterday I got the rejection email. I kind of expected it, because O'Reilly doesn't seem to like me these days, or whatever -- I don't know and it's not important.

Then someone sent me a pointer to http://opensourcebridge.org/ which is in Portland on June 17-19. Now I have an incentive to see if people want to go there. San Jose is closer to Berkeley, so I'd rather go there, but a really open OSCON would be something that's worth supporting. There are other new projects that don't have space at OSCON, so maybe we could all get together in Portland and see what happens.

tinyarro.ws

Wed, 11 Mar 2009 17:16:50 GMT

A new URL shortener.

http://tinyarro.ws/

Ho hum.

No -- it's interesting -- seriously.

The URLs are weird. Are they shorter? Not sure.

The web used to be full of weird ideas that stretched your mind.

Not so much these days.

This is one of them.

Let me know what you think!!

PS: Wish I had thought of it.

Update on the 1000HE

Tue, 10 Mar 2009 23:59:21 GMT

A quick note -- I decided yesterday, that rather than deal with restoring the new netbook that was crippled by malware after just a few hours use, I would instead take advantage of Amazon's generous return policy. I sent the computer back yesterday, and today, before noon, a new one arrived. I spent about half of the day provisioning it, something I'm getting good at.

But this time I've used the new knowledge I have about protection to install various tools that help guard against a reinfection. I promised I'd list those here so others can benefit from the incredible outpouring of information from the members of the community.

First, I declined to update Java and then went to Add-Remove Programs, and took Java out of the system entirely. Perhaps someday this computer will need Java, then I can download it from Sun's website. Until then I don't want to take the chance that it was the opening that the malware got in through. (It's the one thing I updated, and when I went to a perf test site to measure the speed of my connection, the little app was running in Java. It's all I could think of, so it maybe unfair to blame Java, just want to say that.)

Then I installed the following apps: Ad-Aware, Avast, Spybot and Malwarebytes. The last is, according to Stan Krute, a tool that will help if the computer gets infected. Yesterday I learned how important that is. The malware made it impossible for me to get to the Norton site to get a tool that might remove the bad stuff. Same with McAfee and the Microsoft site for defending against malware. It's a good idea to have the removal software already in place when you need it.

Ad-Aware and Spybot are two old friends. In the days of Kazaa (a supposedly nice program that totally ruined a laptop with spyware) they helped get rid of the infections that kept coming back. In those days I was using IE. One of the blessings of this age is Firefox. It may not be the perfect browser, but it isn't full of all the openings of ActiveX and whatever else IE leaves open that the bad guys take advantage of. I will never ever under any circumstances run MSIE again.

Everyone says great things about Avast. I've run it once, it installed easily.

At this point after just a few hours, all the tools say the computer is 100 percent clean. I have the OPML editor running, and uTorrent, Firefox, VLC, SlingPlayer, iTunes and not much more. Ready to kicks some butt I hope.

Folks, this is, in no way, open

Tue, 10 Mar 2009 16:06:03 GMT

Today the Guardian announced their "Open Platform," much as the NY Times did a couple of weeks ago. It's even less open than the Times was.

If it were actually open they'd announce it to all developers at the same moment, so we could all try it out at the same time on a level playing field, not give an advance to their favorites. In the press release they talk about developers who got an early look. Fine. It wasn't open then, that's for sure. Is it open now?

Well: "API key approvals will be granted on a very limited basis, so please don't be offended if we fail to reply to you or don't approve your request in the short term. You can be assured, however, that we intend to open the service more widely soon."

Okay, but please don't be offended if I don't apply for one.

You gotta wonder if when they get out of beta their competitors will be able to repurpose their content. My guess is not. And how broadly do they view their competition? And why should anyone have to guess if they're "open."

All this begs the question, because even if they were just publishing RSS feeds (btw, they are), to be competitive in the API business, you have to enable other people to publish on your side of the API. That was the flaw of the Times model too.

I have no idea how these guys got the idea that they could save the news industry by becoming the tech industry; I don't think they can. What's the diff betw what they're doing and just adding more metadata to their feeds?

My guess this is the result of some tech guys doing their best to give the higher-ups what they want. Some market analyst or consultant told them that to survive they need an API, so come hell or high water, an API is what they'll have.

Correct me if I'm wrong please. (And if the past is prologue, the Guardian will attack personally, calling me names, in print, as they've done so many times. Not to say there aren't a number of very nice people at the Guardian these days. Maybe they can moderate the response keeping it professional and impersonal.)

Big hugs, your pal in Berkeley, Dave

Run the OPML Editor on port 80

Tue, 10 Mar 2009 17:12:52 GMT

I always have to reinvent this, every time I want to set up an OPML Editor server on port 80. So now, by posting it here, I can just search for it and I should find the instructions. If you don't understand, don't worry.

It's now incredibly simple to do this.

1. Add this line to opmlStartupCommands.txt:

user.opmlEditor.flServerOnPort80 = true;

2. If the OPML Editor is running, quit and relaunch.

3. As it starts up it asks for your Admin password so it can do a port forward.

You're done!

"curly"

BMW never returned the call

Mon, 09 Mar 2009 19:11:07 GMT

Closing the loop on the issue with the BMW.

1. The car is drying out. Nice weather over the last couple of days helped. I guess we'll find out when it starts raining again if it still leaks.

2. The independent repair service, H&B, did great, non-warranty work. $120.

3. BMW wanted $800 for the same work.

4. BMW of North America never got back to me.

5. Weatherford BMW, the dealer that sold me the car is a real piece of work. They completely lost me as a customer. In the future, when my car needs warranty work, I'll drive to San Rafael.

6. Thanks, as always, to the Scripting News readers -- a fountain of knowledge and good will!

Reminder: Why I switched to Mac in 2005

Mon, 09 Mar 2009 17:49:25 GMT

I switched because I was Mired In Malware.

I got a new EeePC 1000HE last week, and after just a few hours of use, it's infected with a rootkit virus of some kind. Really clever. Spent three hours last night trying to eradicate it, but in the last three or four years, the malware guys have gotten a lot more clever.

Contemplating switching to the Hackintosh flavor of netbook.

Ran Ad-Aware, getting ready to run Spybot. Downloaded Combofix. I'm going to try to resurrect this baby. Also considering doing a fresh install of Windows but that sounds like more work that Leopard. And then you're still using Windows.

You can read all about my trials on Twitter. But this problem is now serious enough to demand its own blog post. I'm going to see the silver lining here, a chance to learn a lot. Albeit stuff I never wanted to learn.

Of course there will be the moralizing and I Told You So's. Thanks in advance for sharing your wisdom. I agree you are superior and wiser and a better person, in every way.

I don't use MSIE. Please no need to tell me not to use IE. I use Firefox.

Interview with Jay Rosen

Sun, 08 Mar 2009 18:57:52 GMT

It's a good idea to check in with Jay on where journalism is at every once in a while, which is what I did this morning. I'm going to try to do these more regularly with people who are on the Friends Of Dave channel, like Jay.

We start off talking about curmudgeons, then on to rebooting journalism, Meet The Press, the broken government, and everything related. Jay is really smart, spends a lot of time thinking about things I really care about. I thought the interview came out great. Hope you all listen. 40 minutes.

http://mp3.morningcoffeenotes.com/rosen09Mar08.mp3

Jay is a professor of journalism at NYU and was my choice as Blogger of the Year for 2008.

Update: Three articles mentioned in the interview.

A follow-up: It might make sense for Jay to offer one or two paragraph critiques of various bits of journalism. For example this story on TechCrunch is interesting, but it might be more believable if we knew who the author was talking to, or why the source wouldn't go on the record.

Handwritten notes from the interview.

Solving the TinyUrl centralization problem

Sun, 08 Mar 2009 01:33:17 GMT

Following up on an earlier bit about centralizing and TinyUrl, this is one of those vexing problems that actually has a solution!

Every web app that produces long urls should provide a built-in url-shortening facility. The user interface would be similar to the one in Google Maps they call "Link To This Page." You click on it, and up pops a box containing an address you can use to point to the page. Screen shot.

But look at the size of the url that Google gives you. It should be short. Why not something like: http:\//goog.us/8uj9oj.

In other words, why doesn't Google have a built-in shortnener?

When blogging software gives you a permalink, it should be short. It's okay to make the user ask for one, why clog up the system with shortened urls no one uses.

Another key point, when they give you a shortened url, it should point back to the software that gave it to you, so the shortened link will be exactly as long-lived as the thing it's pointing to. In other words, the URL shortener wouldn't contribute any extra link rot, to use an old term coined (I believe) by Jakob Nielsen.

It's a mistake, in hindsight, for Twitter to give us TinyUrl urls, because the link depends on two companies and two servers. It would be better if it just depended on one, less likely to break.

Now that URL length has become an issue for users, it might be even better for designers to view URLs as part of site design. Look at the address for the page for the Wii at Amazon. Wouldn't it be easier to find if the address were:

http://amazon.com/wii

Try clicking on it -- it actually works!

Why should a user ever see the longer crappy url?

In other words, url-shortening isn't just for Twitter users, it's for everyone. Maybe most people don't look at the urls, but some do, and maybe more would if they made more sense?

Note: This is a repeat of a piece I wrote in November 2007.

A short step in URL-shorteners

Sun, 08 Mar 2009 00:04:27 GMT

Progress in the art of twittering comes incrementally. This suggestion is a very small increment but one that would make the job of a frequent linker, such as myself, a little easier.

I would like my URL-shortener to grab the title of the page I'm linking to and insert it into the typing box, before the shortened URL.

Suppose for example I wanted to link to a post Doc's site. From that page, I'd click on the bookmarklet in the toolbar of my browser, it would take me to the shortener page, and this is what I'd see:

Then I click Submit, and off I go. This is a step I do manually now. Better if it were automatic!

Any URL-shortener could do it. The first is likely to get my business.

What blogs are for: BMW

Sat, 07 Mar 2009 18:27:39 GMT

One of the reasons everyone should have a blog is that when a company pushes you around, you have a place to post your side of the story, publicly, so future customers have a chance to benefit from your experience. Over the years I have written up experiences with Travelocity, American Airlines, Comcast, a now-defunct ISP, a Bay Area plumber. Now I'd like to tell you about a problem I'm having with BMW.

I bought a new BMW in July of 2007. It's my fourth BMW and I love it. It's powerful and fast, incredibly responsive. I don't drive much, but when I do, it's still a pleasure. That's saying a lot after having a car for almost two years; I still look forward to driving it. BMW makes a fantastic product.

But -- then we've been having all this rain this winter, and it turns out the car leaks. Water is coming into the cabin, the carpets are wet, they're not drying out even though the weather turned nice a few days ago. So I brought it in for service on Wednesday. The dealer said it was my fault the car is leaking, and wanted $800 to fix it. Now this is a car that has a four year warranty that's supposed to cover everything. I've owned a lot of cars over the years, even a rusted-out Wisconsin junker (that I loved anyway) and I've never had a car leak water. I didn't believe for a minute this was my fault. I told them I live on a normal street, not on a hill, with not many trees (but some) and I could check with my neighbors, but I didn't think any of their cars were leaking. He suggested I call BMW of North America customer service to see what they say, and they said the same thing. I should pay for this because it was caused by an "outside influence" (the rain, I guess).

Then at breakfast on Thursday a friend who also has a BMW says Weatherford is notorious for ripping off customers. Once he brought his car in for service, they failed to fix it three times, and each time wanted to charge him for the repair. He paid, cause what are you going to do, they have your car. Meanwhile they were pressuring me to either return the loaner, or agree to the $800 charge. I told them I was waiting for a return call from BMW of North America. (Three days later I still haven't heard from them.) \

So I went back to the dealer, got my car, returned the loaner, got their writeup of the problem (now the estimate was $625), and took it to a local independent BMW repair shop that gets good reviews (deservedly, it turns out). They showed me a BMW-issued service note, from January 2008, explaining that the 5-series has a problem with water leaks.

I scanned and uploaded the service note: p1, p2, p3.

It's so outrageous. They knew the car has this problem, yet they still wanted $800 to fix it.

10:15AM: I have a wet car that smells bad. I have a call into another BMW dealer to see what they want to do about it.

11:15AM: Got a call back. They want to see the service note. I've emailed him a link to this blog post.

How Battlestar Galactica ends (theory)

Sat, 07 Mar 2009 00:06:27 GMT

Galactica has one jump left before it becomes useless as a starship so they jump back to the home planet that was destroyed by the Cylons at the start of the show.

They arrive, and as predicted the ship is a mess.

They send a landing party down to the planet, but when they come back, they die within hours. Autopsies reveal that the planet no longer supports human life. Only one person from the landing party is unharmed, Kara "Starbuck" Thrace. Now they have to figure out why.

They study her blood and discover by accident that it is identical to the little girl who is half Cylon and half human, the one who saved the President's life with her blood. They send the President down to the planet and sure enough she survives. They send the little girl down and she survives too.

I forgot to mention that the planet kills pure-breed Cylons too.

So they now have a plan. Lots of cross-breeding between Cylons and humans forming a new hybrid race. Kara is the new leader, the prototype, the only adult among them who's healthy (the President is still dying and the kid is too young, everyone else who will survive hasn't even been born yet).

Fade out. Battlestar Galactica is over.

The End.

Tech discussions on FriendFeed

Fri, 06 Mar 2009 00:49:30 GMT

Lately there have been some interesting technical discussions on FriendFeed that I'd like to connect with the technical people in the Scripting News community.

1. Yesterday the question came up why designers of web services reinvent serialization formats instead of reusing existing ones. This is the advantage of XML-RPC. A simple set of types, structs and lists, and a huge set of libraries for all languages. You can write cross network apps at a very high level. An interesting discussion followed, it was nice to close this loop.

2. DeWitt Clinton, a programmer at Google who I've been corresponding with, asked a great question, that I was happy to answer: "Dave, if you could go back in time, would you have used JSON instead of XML for RSS, OPML, XML-RPC, etc, had JSON been popularized at the time?" I think some people will be surprised by my answer, which contained a shout-out to Eric Raymond.

3. I mentioned in one of the discussions and should mention here that I'm thinking about doing a successor to XML-RPC, adding OAuth support. There is some interest, when I mentioned it on Twitter last week I heard back from the people working on WordPress saying they were planning something there. Now that's I've successfully tackled OAuth, it seems it would be a small matter (hah) to take another look at RPC. (It would have a new name, as is the deal with frozen formats like RSS and XML-RPC.) It's now 11 years old, it seems that's enough time to take another look.

What about Sy Hersh?

Thu, 05 Mar 2009 19:40:25 GMT

At breakfast this morning, Berkeley friend and former journalist John Feld said we need journalists to do impartial investigations into government corruption. I asked if he knew of any and he said Seymour Hersh. I agreed.

When he comes out with a piece, I stop everything and read it, as do many others.

John is right, what Hersh does is important, so we should consider that a real challenge. How do we pay for the work he does, and others who want to follow in his footsteps.

Isn't academia the place for a person like Hersh? Isn't that what we want our tenured faculty to be doing -- digging for the truth, no matter where it leads or who is offended? That's what academic freedom is all about.

It would also be great if such "academic journalists" could teach a course or graduate seminar to share their process, teach students how to do what he or she does.

I think it would be even better than having them work for big media companies, because then they could go after the BMCs, and lord knows they need going-after.

Jon Stewart reviews CNBC

Thu, 05 Mar 2009 20:56:08 GMT

Maybe Schieffer should Face the nation?

Thu, 05 Mar 2009 02:49:45 GMT

I tried to listen to Sunday's Face The Nation podcast, an interview with Rahm Emanuel, but I couldn't stand it. I've really gotten out of the habit, and now with fresh ears I know exactly what's wrong. I want the interviews to be grounded in the reality that we, the people, live -- not the make-believe logic that governs the ruling class in DC. Schieffer kept asking questions Republicans would ask to try to make trouble, but I understood they were based on an unstated and unproved premise that earmarks are inherently evil. Emanuel was answering the questions directly -- yes we will have earmarks. Schieffer kept playing the gotcha game, but it was stupid, Emanuel had conceded the point! OMG.

Our economy is crumbling, and these guys are arguing nonsense. We have important business to conduct, saving what we have left of our way of life. But you can't have a realistic conversation without some supposed "journalist" trying to trap you into telling a truth you're willing to stipulate.

The solution is simple -- give Schieffer a script written by real people, and if he won't do it, get someone who will. There's no time to screw around. We need to get real, quickly, without any delay.

Now for the good news, sorta. I listened to a fantastic interview by Terry Gross with Simon Johnson, former chief economist of the IMF. It's both sobering and encouraging. He lays out what we need to do, simply and clearly. It's just common sense, highly recommended.

Archiving your tweets in XML

Thu, 05 Mar 2009 01:58:57 GMT

When I mentioned that I had a tool that archives my Twitter posts, and those of people I follow, a fair number of people asked that I release the code. I have done so, the app has a user interface and docs, and if you want to try it out or run it, you're welcome to.

The software runs in the OPML Editor, on Mac or Windows. It maintains a folder of folders of OPML files, one for each user, organized in a calendar structure, one for each day; and it keeps an index, also in OPML, and a weblogs.com-compatible changes.xml file for the whole thing. The pointers in this paragraph point into my archive. If you run the software, you will have your own versions.

I also added a feature that automatically (and optionally) synchronizes the folder with a structure on Amazon S3. I want to encourage people to try out S3. You don't need a lot of technical skills to do it. I've included a section in the Howto that walks you through it.

If all this sounds confusing, start here, and follow the instructions, carefully.

http://editor.opml.org/twitterCalendarTool.html

If you have questions, post them here or as a comment on the Howto.

Good luck and I hope you enjoy it!

PS: There may be some interesting applications that can be built on this structure of folders.

Poor man's email?

Wed, 04 Mar 2009 13:45:59 GMT

One of my favorite blogsports for the last couple of years is pondering what Twitter is. Here are some of the things I've come up with:

1. Personal notepad.

2. Coral reef.

3. Publishing platform.

4. River of news.

I'm sure there are others, and as I think of them I'll add to the list.

But one thing I never thought of Twitter as was Poor Man's Email, which is how Google CEO Eric Schmidt described it to analysts yesterday.

My first inclination was to shout out something about Schmidt, but I held on to it, instead deciding to give it some thought and let other people go first. Surprisingly, there hasn't been much reaction.

Schmidt has a PhD, and a long track record in the industry before he went to Google. I have a postulate that if very few people do something then it must be hard, and therefore whoever is doing it must be smart. So when people say Steve Ballmer is dumb, I don't buy it. Same with Schmidt. So I wonder how calculated the statement is, or if I'm missing something -- because it never, ever occurred to me that Twitter was any kind of email, rich man's, poor man's or middle class.

I couldn't imagine two things being more different, Twitter and email.

1. Twitter is primarily one-to-many, where email is primarily one-to-one.

2. Twitter is by default public, where email is by default private.

3. Could you use email to implement something Twitter-like? Yes. Could you use Twitter to implement something email-like? Yes. But neither is the same as the other.

Schmidt talks about software the way I think about it. Once you have a base set of features it's an interesting puzzle to decide how to evolve it. There's no doubt Twitter has a tricky evolution in front of it. No matter what it does, it's likely to upset users, just as every Facebook move inspires an uprising. If Twitter had established a history of quick feature upgrades it would be a different story, but there were no new features in 2008, and so far none in 2009. That's a long time between changes. At some point they're going to add new features, if only to keep even with the competition that is sure to come. What they choose to do will set expectations for what's to come. The longer they wait, the harder it becomes.

So Schmidt laid it out.

The key question is -- does the basic unit of Twitter change and if so, how? Is the 140-character limit sacred (my guess is yes). What metadata will accompany a twit? This is where it gets interesting. It's hard to imagine Twitter passing on the temptation to add location data to each message. What about the URL? Would they consider moving the link out of the 140 character space and making it part of the package? What about incorporating "re-tweets" into the architecture, instead of forcing users to invent new language, and using up another 20 characters of the 140 (by convention). They could make a lot of improvements if they added more structure to a tweet. I'm sure there are people of both minds inside the Twitter company.

Another takeaway from this is that Google is watching. If Google is preparing their own Twitter, what will it look like? Will it have a different, incompatible API? (My guess, probably. Google will make their own play for developers.) Will it have the same limits as Twitter? (Probably not. This is a very easy way to put pressure on Twitter, even when they have the installed base advantage. Users will say "If Google can do it, why can't Twitter?")

In Marketing Warfare, Ries & Trout tell the story of how Pepsi got a slice of Coke's market in the 1930s based on bottle size (page 119). Coke had a huge installed base of machines that could only serve 6.5-ounce nickel bottles, which they thought of as the perfect size for a cola drink. Pepsi thought 12-ounce bottles were better, so they came at Coke with the larger bottle. Coke was wrong, but it took a long time to figure it out. Eventually they threw out their machines and bottles and matched Pepsi.

Pepsi-Cola hits the spot. Twelve full ounces, that's a lot. Twice as much for a nickel, too. Pepsi-Cola is the drink for you!

This is why Google is likely to have a 160 character limit. (People who said Schmidt was in error didn't consider the possibility that it was a deliberate misstatement.) And it seems likely Google's Twitter will be based on GMail. (Paul Bucheit, a founder of FriendFeed, could probably comment on the likelihood of this working.)

Finally, I couldn't help but notice how similar Google's reaction to Twitter is to the reaction of all market leaders to initially successful upstarts. It's the same reaction Alta Vista and Yahoo had to Google when they were young. "Search is only part of a portal," they sniffed. You can make a hugely long list of BigCo's that failed to understand the threat presented by upstarts, and lived to regret it. It's hard to think of a a single example of a BigCo that took a threat seriously when it (based on historical hindsight) needed to. Maybe this is because people like Schmidt, while they are educated, and intelligent, weren't there when their company was the upstart, and neither were most of the people there now. The institutional memory fades, and as it does, it creates the opportunity for the next generation? Perhaps. We'll get a chance to find out soon enough.

Fractional Horsepower Twitters?

Tue, 03 Mar 2009 21:59:03 GMT

Fractional Horsepower is a very powerful idea. It says that sometimes you can make a new product by taking an old one and scaling it down.

The originator of the concept, in my experience (he may have borrowed it from someone else) is Steve Jobs, who described the Apple II as a fractional horsepower computer. In those days computers were big, Jobs believed and many of us agreed that a lot could be gained from taking the big idea and making it small.

If you look at the history of computers and publishing (the two things I care most about, professionally) you can see that the trend is going, inexorably, that way. Things keep getting smaller, and every time they do, huge power is unleashed. Maybe it's like nuclear fission, there's this huge power holding the nucleus of an atom together, making all those protons stick together, then it's unleashed, boom (another Jobsism) a big explosion.

We like netbooks because they're smaller than laptops. We like iPods because they do what stereos do but fit in a pocket.

I wrote, many times about Fractional Horsepower HTTP Servers, and today they're a reality. Every device that can be configured through a browser has a little HTTP server in it. Each of them has a single user, they sit idling most of the time waiting for you to do something. My printer has one, my receiver has one. Look around, they're everywhere.

The other day I wrote a piece called A billion Twitters. In this environment a lot of people just skim, and I think they didn't read it because it seemed to be saying that there would be billions of Twitter users. I don't doubt this, but that's not what the piece was saying.

I am pretty sure the same logic that led us to personal computers will lead us, inevitably, to personal Twitters. Yes, there are huge advantages to scaling up, not down -- and that was true in the earlier shifts too. We loved our Apple IIs, but banks and airlines needed massive computer resources to do book-keeping and reservations. And we love our search engines, and web apps, all of which are made possible by scaling up.

As a thought exercise, I tried to imagine places I would put a Twitter if I had the power to do so. I would certainly put one here on my website, to enable Twitter-like micro-publishing among members of the community, in a sense to define what it means to be a member of this community. I don't imagine either the blog or the comments going away, in fact I am sure they would be enhanced by our own Twitter. We could try to organize a community on the main Twitter, but the comments that are relevant to this community would scroll off far too fast.

I suggested to Craig Newmark that he consider adding Twitter-like functionality to Craig's List. He asked what that would mean. I said I didn't know. It was part of the thought exercise. I posed the question on Twitter this morning and got back a huge number of comments. I wonder if the same will happen here? We'll see.

It's hard to imagine that Twitter is so unlike everything that came before that it won't go both ways as every other publishing technology has -- both up and down. I'd like to try putting a Twitter on my netbook and see what happens. Probably nothing, but you never know! That's how creativity works, play what-if and relax all the constraints and challenge your mind to make sense of it. Most things never do make sense, but every so often there's a winner. RSS was such a thing, as were blogs and podcasts. What if there were an XML rendering of this blog? What if everyone had their own website? What if radio didn't require air waves? What if everyone could have their own Twitter?