Thursday, January 31, 2002

New Scientist goes Copyleft The british popular science journal, New Scientist, published an article on copyleft under a copyleft license. The question is why?

Publicity and having a cute gimmick was obviously the key reasons--the article appeared on slashdot, which is always nice for traffic. But the value of text online is that people can read it and comment on it freely, just like I'm doing here. I don't want to edit the text, just point at it and pontificate.

Under a DRM world, the article would have to be under password protection, which meant I might not be able to point at it (the way I can't point to Wall Street Journal articles, and try not to point to NYTimes articles). Also, I can cut-and-paste quotes into this post right off the screen, which I could not do under CD crippleware style regimes. And the article does not need to have distributed hosting, it just needs to keep a static URL to be 100% useful.

The article dismissively notes "It's hard to avoid the conclusion that Open Audio amounts to little more than an opportunity for obscure artists to put themselves in the shop window" ignoring the fact that getting in that shop window is incredibly important to artists, and is controlled by distributors. This is why any studio can take anyone off the street and make them a star. And since they originally signed up cheap, they have to produce multiple albums for peanuts. Anything that threatens control over the delivery channel reduces the value of this resource. The RIAA does not like bargaining with performers who have their own distribution channel and audience.

By contrast, copyleft is designed to keep working code available and useful, retain fair use, and prevent collusion that artificially restricts quantity and raises price to earn monopoly profits. Text online doesn't need copyleft, it's all essentially released into the public domain anyway.

Wednesday, January 30, 2002

Struggles with Radio Lawrence Lee, who works at Userland and maintains the incomparably excellent Tomalak's Realm heard about my troubles with Radio and suggested I made sure "" was running on a browser after I started the system. This actually worked (thanks Lawrence!) but note how this problem KILLS the customer experience. If, for some reason, a browser pointed to that URL does not appear after starting Radio then the product appears utterly useless for that user. It's very tough for Userland to recover from that error and keep the trial user. Others have emailed me reporting similar experiences.

But since Lawrence was kind enough to help me out, I did recover and got things started, even posting a few notes on my new Radio site. All well and good. But the problem was--the site sucked. The typical radio template ("theme") is design intensive, right down to the 1 pixel gifs for spacing. Slow slow slow over dialup, hard to edit, and causes Netscape to hang on my system. But when I tried to pare things down so they more closely matched this site it was like pulling teeth. Reams of convoluted table within table HTML, tags ("macros" maybe?) that I couldn't understand, four different levels of templates which I couldn't tell apart, and a discussion forum (for help) that lacked search, making it hard to find specific information. Bad customer experience.

So I used google to crawl through Userland's discussion groups, found some helpful stuff, and dinked around trying to slim the site. But I was doing all this in a web application and wondered "what does the download do?" There was no obvious documentation, and I still have no idea. Infact, I have no idea what 99% of the stuff in Radio does.

Why am I spending time on this when Blogger serves my needs so well?

Blogger is wonderful--easy, fast, simple. But the service has been giving me some problems. Sometimes it's down, and it'd be nice to have something useful run on my desktop for those moments. It recently ate my templates (for no good reason). It recently had issues with ftp'ing stuff, even though everything worked fine before. I hoped Userland would give me Blogger ease-of-use with more reliability and was willing to pay for that. But that didn't happen so I'll check out Blogger Pro instead.

I'm clearly not the only one considering switching from Blogger to Radio. I'm sure Dave and Ev don't want to compete fiercely, but it wouldn't be terrible if Radio (much smaller user base) kept this in mind and anticipated the needs of switchers. They could make it easy to move archives over, make it easy to move templates over, and make it easy for folks to understand how Radio works given that they know how Blogger works. Instead, Radio gives us Blogger API in Radio which I'm sure is wonderful, but I have no idea what it is and it doesn't help me get my blog working.

I like Userland. I like Dave. I wanted to like Radio, but I don't understand it. It's too complicated. It doesn't help me get my work done. It makes it hard for me to switch. Maybe 9.0 will be better. Maybe I'll try again then. Maybe not.

Tuesday, January 29, 2002

Fair use vs. DMCA As predicted, fair use will kill (change) the DMCA. Because "fair use" is not a right--it merely limits the copyright holder's ability to enforce copyright--DRM systems eliminate it complete and legally under the DMCA. Obviously this has to change, or as Boucher puts it: "The only conduct that should be declared criminal is circumvention for the purpose of infringing a copyright." But enabling non-copyright infringing uses of digital copyrighted material would make a DRM system even more hideously unworkable than they are now. This will further raise the price of copyright enforcement, tilting cost advantage towards systems that embrace public domain information. I see such interests pushing for expanding the public domain over the next ten years through reducing copyright's ridiculously long term.

System failure It's impossible to predict short-term stock fluctuations, but the Times is calling today's sell off the result of public doubt over accounting quality. Earnings quality is one of those issues that keeps reoccuring, and it's no different with Enron (this isn't to say accounting standards shouldn't be improved, I just wouldn't hold my breath).

Faster Macs Apple released new G4 towers which finally break the 1 Ghz barrier. But as critics point out, processor speed isn't everything and Apple's specs are still lumbered with old system memory, multi-processors, and lashings of cache. They have not overcome their lousy chipsets. But again, processor speed isn't everything. Unless you actually need power-hungry applications, you'll do fine using light equivelents that work faster for less system strain than the bloated Office suite. Engineers don't appreciate how the experience of speed can be uncoupled from processor cycles. The Good Easy is one example of how this can happen, Palm is another.

Monday, January 28, 2002

HP vs. HP The family fued between Hewlett Jr and Carly is essentially an argument over how to deal with PC commoditization. Microsoft (and Intel) have made PCs utterly interchangable, with profits only available to lowest cost assemblers (like Dell). Hewlett Jr's strategy is to drop this business like the dog it is, while Carly wants to hang onto value by betting on Microsoft IIS related services. The former doesn't destroy value, but doesn't deliver growth either. The latter will probably do neither. If HP is interested in services, it needs to abandon a platform strategy (which IIS emphatically is) and become like IBM. If it wants to stay a tech leader, it needs to commoditize complementary products (like operating systems) and keep economic rent in hardware, not give all profits over to Redmond.

Friday, January 25, 2002

Nostaligia The BBC Micro Model B was the very first computer I owned (made by Acorn). I taught myself how to program in BBC Basic, which I thought was lovely. I thought the platform had gone forever, but it lives as a *Nix Interpreter. Joy!

Markets for GNU/Linux While IBM reveals its first all GNU/Linux mainframe, Loki, the GNU/Linux game company, has shut its doors. It's not surprising to see why. IBM's old mainframe OS, z/OS, was just something they had to produce to sell their big iron. Since they now can get out of the mainframe OS business, they should, just as Apple has exited the kernal business. Total cost of ownership for their machines goes down, R&D costs for the OS go down, and everyone is happy.

In contrast, Loki was trying to sell into a market that doesn't really exist. It's a trivial observation: most people use Windows, and even those who use GNU/Linux can probably dual-boot to Windows also. So why write games for any platform other than Windows? Loki did it for love, but still thinks it could do it for money.

Thursday, January 24, 2002

Price discrimination Price discrimination helps companies sell their product more efficiently (maximizing surplus) and shifts surplus from consumers to producers (something polite societies officially frown upon). In essence, if I can sell some of my stuff for $15 and some of it for $5, more people can get my stuff than if I could only sell at $10. But it also means that some people are paying $15 for what they could have gotten for $10. It's not clear exactly how much legal protection companies have for price discrimination. They can offer paperbacks several months after hardbacks, but can't stop you selling your book after you've read it. In the UK, however, it seems that regional encoding (a way of price discrimination by geography) is protected and enforceable by law.

Wednesday, January 23, 2002

Couldn't figure out Radio 8 I'm a fan of Userland, and I was genuinely excited to try out Radio 8 (their new web publishing tool) but I was utterly incapable of using it. I downloaded it, fired it up, and then--nothing. I hit "New" and got a command prompt. I searched the web for how to "get started"--nothing. And this is for a service that is meant to be easy to use. I guess I just must be dumb. Thank goodness for Blogger, which genuinely makes it really easy to start a simple weblog. It's sad when a company as dedicated to creating good customer experience as I know Userland is makes a tool so difficult to pick up and use.

Tuesday, January 22, 2002

AOL will not buy Red Hat Despite all the rampant speculation, I think it's pretty clear that AOL will not buy Red Hat. Why should they? Red Hat has done a pretty good job of not differentiating itself in any way, nor does it bring any scarce resources or defensible competitive advantage that AOL might be interested in. Moreover, the only valuable resource Red Hat really does have (engineering talent) is mobile and could simply be poached by AOL for a fraction of an acquisition price. So, there may be defections from Red Hat to AOL, but I'd be surprised by any purchase.

Monday, January 21, 2002

Law and Economics It's frightening how poorly lawyers (and legislators) comprehend the economics behind the laws they pass. This article on "lockware" (which I call crippleware) talks about how DRM allows companies to increase profit through price discrimination. Price discrimination may or may not make markets more efficient (depending on how efficient they are currently) but it will move surplus from consumers to producers. Perfect price discrimination shifts ALL consumer surplus to producers. The US (and many other countries) have taken a pro-consumer stance on this issue, and mechanism that shifts surplus in this way is contrary to that position. Now, how efficient are markets for information goods (like music)? Terrible. The extraordinary length of copyright essentially gives the recording cartel unlimited ability to extract monopoly rents through artificially restricting quantity. The cartel also sets prices high through collusion. Because of this, I don't have many sellers competing to offer me "The White Album" even though half the group are dead. DRM might make this market more efficient, but it will also shift what meager surplus consumers have to producers. But this won't help artists, because in the music biz, distributors take most of the surplus (their quantity restriction acts as a "tax") which also shrinks the market for both consumers and producers (musicians). Distribution of live music also has a huge tax, Ticketmaster, whose "inconvenience charge" can add 50% to the cost of a ticket. Given there horrible inefficiencies, ruinous taxes, and shrinking value (surplus minus price) for consumers, it's not surprising folks aren't buying much music anymore.

Saturday, January 19, 2002

Housekeeping I'm shifting my host from the excellent to It's not fair to keep nagging Alan to get the site up when he's already being so generous in volunteering his time. Also, I'm shocked by how badly my site renders in Opera, so am trying to clean up the HTML behind it. Wish me luck. Expect outages.

Friday, January 18, 2002

Stallman goes to Chicago Richard Stallman is the Mahatma Ghandi of the software world. Long before the dotcom boom came and went, he was writing Free Software at MIT. Known within hacker circles simply as RMS, Stallman believes that people should not be prisoners of their technology, and so set about writing the GNU operating system (which now surrounds Torvald's Linux kernel to make GNU/Linux), founding the Free Software Foundation, and drafting a license so unusual Microsoft calls it a "cancer." On Halloween 2001, RMS came to the University of Chicago to speak about freedom and his vision for the future of software.

RMS's goal is entirely ideological. He's on a mission to liberate all people from the prison of proprietary software. He shuns material possessions and lives like a student so he can pursue this goal unencumbered. While Torvalds may shun the "Ghandi" moniker, Stallman, with his uncompromising ideology and ascetic lifestyle, is a strong candidate.

Stallman characterizes Free Software as software which gives the user the following "four freedoms":
  • The freedom to run the program, for any purpose (freedom 0).
  • The freedom to study how the program works, and adapt it to your needs (freedom 1). Access to the source code is a precondition for this.
  • The freedom to redistribute copies so you can help your neighbor (freedom 2).
  • The freedom to improve the program, and release your improvements to the public, so that the whole community benefits. (freedom 3). Access to the source code is a precondition for this.

He views these freedoms to be as inalienable as the right to self determination. Therefore, using proprietary software that does not give users these freedoms is morally equivalent to selling yourself into slavery. This is the single schism that divides the Open Source and Free Software movements. Open Source thinks that Free software is good and desireable, but that Stallman's four freedoms are alienable, and it's not wrong to trade them (by using proprietary software) for other goods.

RMS enshrined these four freedoms in the GNU GPL (General Purpose License), which is as much political manifesto as it is a software license. Stallman is not shy about admitting this, Free Software in his eyes is solely a moral (not commercial) concern.

Stallman views the four freedoms as being so sacrosanct that it is better to go without software than to use proprietary software.

Most people don't share RMS's position. Most people are OK with giving up (some) of the four freedoms if it means they can have software. If proprietary software delivers more value (utility minus price) than free software, people are OK with using it.

Which brings us back to the "freedom vs. power" and "open source vs. free software" debates. The question fundamentally comes down to: would you prefer the option to buy proprietary software if free software does not a) do the job, or b) exist? If you believe Stallman's four freedoms are inalienable, the answer is "no". If you believe Stallman's four freedoms are alienable, the answer is "yes".

So is this a critique of the GPL? Not really--the GPL is the best protection the technology community has against things like patents and rabid copyright. But should the GPL be mandatory for all code? Clearly not, as most people do not believe Stallman's four freedoms should be inalienable.

This posture also limits tactics the Free Software Foundation could take to bolster the fight against legislation as oppressive as proprietary software, but beyond the scope of copyleft. The EFF needs money to fight the DMCA, SSSCA, copyright extension, software patents, mandatory DRM, and a host of other moves designed to extend monopoly and suck all innovation out of the technology field. Nothing in the GPL excludes dual licensing, which would generate funds the EFF could use and increase the amount of Free Software in the world, but FSF's philosophical posture forbids them from doing this. That's a shame.
Link to this article

Thursday, January 17, 2002

Handspring exist Palm market Dubinsky announced that Handspring was looking to get out of the PDA market--good for her. The PDA market is glutted with over capacity that, especially since Sony is in the fray, will not leave. Moreover, the struggling Palm has been very ambivelent about its OS lisensing policy--it's not clear what they will do in the future. And at the high-end, Microsoft is handing out $400 with each WinCE unit it sells, so I don't see any profits appearing in this industry anytime soon. Moreover, I don't see much revenue here either. Compared to the cell phone, the PDA market is finite, with paper providing a reasonable substitute for many people. Hardware prices, while falling, aren't going to reach $50 / unit anytime soon, and so aren't going to drop to the level where a whole new market of folks will appear who would have bought one but couldn't afford it before. And finally, Palms are just too good. I bought a Visor about two years ago and there is no reason for me to buy another one until this one croaks. It does everything I want about as well as can be expected in a world without magic. The biggest hinderance to my PDA experience is that it conforms to Outlook specifications, which suck. So, slow upgrades, no profits, saturated market. Not a good place to be.

Wednesday, January 16, 2002

Useless PDA, USELESS Argggh! This article infuriates me. It's about how Microsoft "understands" PDAs and is going to dominate the market because: it's bundled WinCE into .NET, installed I.E. 5.5 as the browser, put in Windows multimedia player, Passport, and some unidentified networking standard. Let's look at this from the perspective of a normal human being: I don't care about .NET, there's no network for my PDA to connect to, web pages look lousy on a tiny screen, movies look lousy on a tiny screen, I don't need Passport, and did I mention there's no network for my PDA to connect to? The only thing that shifts WinCE devices is the $400 Microsoft shells out with each one.

License 6.0 License 6.0 removes the only competitor Microsoft has left--older versions of Microsoft. Under it, software will automatically expire unless the benighted user continues to pay Redmond. Touted as "software as service" it's really just "software as rental," designed to extend the monopoly rent Microsoft gets on its commodity Office applications. Remarkably, CIOs have fought back--which is the first step in improving the productivity people get from computers. If you don't take control of your software, it'll take control of you, and upgrading as you see fit is a very basic sort of control. More interestingly, Microsoft is offering discounts to customers who then agree to never consider a competitors product. Anti-trust anyone?

Copyright in New Yorker You know obscure technical battles are going overground when something as arcane as copyright and technology reaches the pages of the New Yorker. James Surowiecki talks about how Disney, and various other copyright interests, have extended its length to the criminally long lifetime plus 70 years.

Here's my call: I think that content industries will get more laws like the DMCA passed and have more control over copyrighted material online. But I think it will cost them the length of copyright. Moreover, I think that fair use will be eliminated in digital media (i.e. never be allowed in the first place), but I think this too will contribute to copyright's term being further limited. (PS Sorry about all the site outages--I'm working on fixing them)

Tuesday, January 15, 2002

Opera I downloaded Opera for Mac yesterday and now need to decide whether or not it's a good replacement for Netscape 4. Advantages of Opera: does not hang on Scripting, advantages of Netscape: new windows open at homepage, and handles text fields better.

Monday, January 14, 2002

Philips audio vs. Philips video Companies can be delightfully schizophrenic. Just when I note how Philips CD player makers don't want DRM CD's, an article comes out detailing how Philips TV makers want wireless DRM for video. Consumers will hate this and the products won't sell, but Hollywood has obviously put out an RFP for systems that keep digital content from being locally redistributed. This is good for Hollywood, who by restricting quantity protect the profits they get from movies now, but bad for consumer electronics manufacturers, who can't make devices that consumers want. After a few market failures, they'll figure this out. Consumer electronics (and PCs) need new product categories badly.

Friday, January 11, 2002

Conflicted interests Now that @Home is no more, AT&T provides both internet service provision and pipes, resulting in shoddy service and forcing customers to use their products, even if those products are inferior. Similarly, companies that own content and distribution equipment (Sony, AOL) are conflicted over DRM. The content arm wants the ideas locked up right, while the distribution arm recognizes this will make their products fail in the market. Philips, who makes CD players but doesn't own movie studios, acts rationally and defends its own self interests by arguing against DRM and working to commoditize the information goods complementary to its hardware. The market itself will not hesitate in rejecting DRM.

Housekeeping I recently asked whether people were finding this site hard to read, and got the following excellent responses. Mark Pilgrim writes:
No problems with Mozilla 0.97 on Windows. The person having problems may have been on Linux. I had no end of font problems using Mozilla under Linux, and eventually gave up and set it to use specific fonts (that looked good) and ignore the fonts specified on all web pages.

Jubal Kessler adds:
Tell that mozilla user to install Verdana and Arial truetype fonts from Microsoft's font page, and point Mozilla at them instead.

It's what I'm doing on this RedHat 7.2 laptop. THere may be logistical issues with extracting the fonts from the .exe on Linux, but the workaround is to get a friend to zip the Fonts dir in Windows and send it along.

I don't have a problem reading your site since the Verdana font kicks ass.

This should solve all problems, site stays the same.

Thursday, January 10, 2002

Comments? I slapped together this website for the University of Chicago, Graduate School of Business' High Tech Group, and it was taken down by the director of Web development because (s)he felt it was inappropriate. And they could well be right. So what do you readers think? As a potential student or employer, would the High-Tech site turn you off University of Chicago MBAs for potential hires? Email me at with your honest opinions. Thanks!

Everything restored Just restored all the archives and all the links. One relearns to back up the painful way.

Response to Mac speed problems USS Clueless writes an excellent reply to my post on how Mac's can tackle their Mhz problems. He's right in pointing out that Motorola's chip will break 1Ghz eventually, but that it's rate of improvement is lower than Intel/AMD. He's also right in pointing out that software is hard, and writing large, complicated packages, is harder still. Finally, he is also correct in saying a zippier program that's much more expensive will lose to a cheaper program that's out sooner and less fast. The market's spoken eloquently on that many times. But with all the recent Be nostaligia, I started wondering why other OSes can't use hardware as well as Be seemed to. While Apple clearly has no long term incentive to write particularly efficient code (they sell hardware), it's a response they could have used when the PPC stalled.

Video Napster I have no idea why the movie folks are stressed about video being Napsterized--they clearly have never sat at home and tried to download a video. Over a dialup connection, downloading a video is a horribly slow experience, so horribly slow, infact, that it might as well be impossible. Over broadband it's less awful, but still lousy compared to visiting Blockbusters. And besides, no one has broadband and no one is going to have broadband anytime soon. Instead of of joining the RIAA as most-hated industry group, they should just do nothing and discover they're fine.

Link problems Well, all the posts are back up but the "links", "essays" "longer pieces" etc. files are all gone. I'll have to redo them :(

Wednesday, January 09, 2002

Apologies Sorry for the website going down. Is back up now. Thanks Alan!

Tuesday, January 08, 2002

Macs are still too slow On the subject of Motorola's incompetance, Apple still hasn't broken the 1Ghz barrier. This article goes through good reasons why Apple screwed up, and why there's no light at the end of the tunnel PPC speed wise. But Apple is correct in saying Gigahertz aren't everything, even though not for the reasons they like. In this post-BeOS world, one has to wonder why other software doesn't do a better job optimizing software for hardware. Apple could easily improve the speed experience of it's machines by writing better software AND/OR using multi-processor CPUs. Normally, Apple's not interested in optimizing software speed because it wants to sell faster computers, but given it's current embarrasing sub-1Ghz position, and the speed shortcomings of OS X, nimbler software would help. Moreover, an OS that can use multi-processors well (like *Nix) would find a happy home in multi-processor boxes, something that Apple should like given how Motorola (and IBM) don't seem interested in producting anything much faster than the G3. It would even give Apple a new box to sell.

Customer driven design The Hawkins, Dubinsky team that brought us the original Palm stand out as hardware folks who truly design products around peoples' needs. This article talks about how they brought in ethnographers to examine how people use cellphones "in the wild" and lived with prototypes and foam mock-ups when designing the new, wonderful Treo. Contrast with Motorola's excrutiation: the V200 Personal Communicator. This horrible device was clearly concocted in some dank R&D lab, never coming into contact with a normal human being, or, one suspects, the light of day. Treo may do quite well (high price tag is a strike against it), V200 communicator will fail.

Monday, January 07, 2002

Macworld There is only one topic to blog today: MacWorld. As expected (and leaked by Time Canada) Jobs unveiled the new iMac. It looks like a desklamp, and I fear it will fare as badly as the Cube. But I'll show it to some normal people and see what they think. The iLux (as it's already being called) has impressive specs and I'm glad to see the updated keyboard has extra function keys -- powerup Keyboard Maestro and start building the Goodeasy on OS X.

iPhoto on the other hand, seems unreservedly wonderful. I was on the verge of selling my rarely used FinePix 2400Z digital camera because the awful accompanying software rendered the camera useless. But with iPhoto, it might be insanely great yet. One more reason to upgrade to OS X.

And on the subject of OS X, the pressure to switch continues to build. I beleive OS X is the most important Apple release in the past five years. Expanding the Mac development community by embracing Unix coders, expanding the number of Mac apps by embracing POSIX compliant code, and expanding the technical resources available for kernal development by embracing open source, BSD, and Darwin, dramatically improves Apple's competitive position vis a vis Microsoft. Sadly, fear of the Dock, fear of losing my Goodeasy speed, and fear of shelling out $500 for Office keep me from switching. My Mac is so great now, it's hard to switch to something that has had mixed usability reviews in the press. Maybe OS X.2.

Check out this wonderful keynote blog (hosted by Userland.

Friday, January 04, 2002

Housekeeping Some good news: winterspeak was picked as a blog of note--very exciting! Thanks Ev! (Ev created the wonderful blogger that powers this site). Also, I'm trying out permanent links, so we'll see how they work (I'll see if I can update the archives with them also). Finally, let me know if you're having problems reading this site. A Mozilla user having trouble with 12 pt arial requested I remove all font specifications, but I'm leery of doing this. Email me at

Public Money, Private Code My friend EV sent this story on how universities, obsessed with licensing code for money, are forbidding professors to release their work under Open Source licenses. This is particularly worrying because of the role Berkeley's BSD in creating the Internet.

I don't see it as quite the catastrophe the article portrays. Before licensing, much potentially useful research never went anywhere because the producer had no incentive to take it to market. Now technology can be licensed, so producers are more vigilant about making sure useful things get to people, which helps everyone.

There are some problems. Firstly there's the agency problem--the university owns the IP and its interests are not the same as the researchers. A researcher might want her idea to be widely used, but the university wants to restrict the quantity and so make excess profit. Secondly, publicly funded "blue-sky" research plays a critical role in furthering science, and keeping an eye on commercial viability may stifle this. I would correct both of these by granting the IP to the principal researcher, who can then choose to sell it to the University, to a company, or to the public domain. Moreover, a smart researcher (and university) could do both under a dual-licensing scheme (GPL + commercial license). This way the product can be taken to market but is also free for everyone to use and build on.

Thursday, January 03, 2002

How to kill a format The new Napster service will fail because of the restrictive .NAP file format. The recording industry would like to kill MP3 because it's insecure, and techno-libertarians would like to kill it because of onerous patent restrictions. The latter will struggle because MP3s feel free to users. The former will fail unless Microsoft and the US Govt get together and force a change through. Better compression and sound is not a factor because MP3s are good enough. Unrestricted use improves the experience and gives such formats a competitive edge. And MP3 already enjoys demand side network effects.

On the subject of Microsoft vs DoJ, only 22 days left to file your opinion.

Wednesday, January 02, 2002

The Joy of local storage XMRadio beams 100 radio channels to cars kitted out with special expensive gear (oh, plus $10 a month per car). Sounds like anything customers would be interested in? I didn't think so either. It's surprising to me that companies are still banging their head against the last-mile bandwidth problem when it's patently clear this sector will remaind stagnant for at least five years. Storage, on the other hand, continues to improve at a rate that would make Moore blush. TiVo uses local storage to deliver a great viewing experience even though video isn't being zapped to it over IP via a fibre optic network. Apple's iPod delivers celestial jukebox in 5 gig slices without a wireless internet connection in sight. If I plug my iPod into a car stereo (through a tape deck if I must), I essentially have what XMRadio is offering, but better, for cheaper. Good innovations can come out of massive micro-storage now, broadband isn't coming anytime soon.

Commoditizing Complementary Good Dave Winer seems to have settled an ongoing feud against Open Source saying that it's sometimes OK to give away code. Sometimes, it's very OK to give away code--here's my response:

Hey Dave:

Nice email. I know Userland has been upset about Open Source, but it really shouldn't be.

> Simply stated, here's my rule. Where you want
> competition, give away the technology. Where you
> want to be competitive, keep it to yourself.

Also known as "commoditize away the complementary good" (Shapiro and Varian write about this a lot in "Information Rules.") Microsoft did this with hardware, reducing Intel-PC components to their average variable cost. When you do this for silicon, the price falls to pennies, but when you do it for software, the price falls to zero.

Hardware vendors should like Open Source--it makes their products cheaper and more profitable. You should expect hardware vendors (from TiVo to IBM) to continue offering Open Source software, because they want code to be commoditized and earn margin on hardware. And remember, Open Source embraces interoperability in a way closed source software (free or otherwise) often doesn't.

Also, Open Source does not have a great track record when it comes to user friendly applications, because ease-of-use may not be interesting to most developers. But this is where companies like Userland, who care greatly about ease-of-use, can find valuable markets. And if Userland can't come up with better, easier-to-use applications than a bunch of coders tapping away in their free-time, does it really deserve to exist? But I have faith in you guys, I'm sure you'll turn out great stuff worth the money.

Open Source is as much a reaction against lock-in and broken interoperability as anything else. Most users are not like ESR and would welcome quality closed-source applications (even on Open Source platforms) that are materially better than Open Source equivalents to justify their cost. Especially if they are innovative, don't lock customers in, and play nicely with other programs. Sounds like a good place for independent software developers to live.

Apple's BSD kernel is exactly this thinking in action. Aqua, for all it's problems, provides a richer GUI and application environment than GNOME or KDE. I'm glad Apple stepped in to fill that void and bring Unix plus GUI to market (now if only it's GUI acted more like Unix). Mac OS X offers more stuff than OS 9, including a rich development environment, the entire universe of POSIX code (almost), and an Apache server. Contrast this with XP, which offers less functionality than previous versions of Windows (product activation, hardware binding, draconian licenses).