Sunday, March 31, 2002

Clear Channel vs FCC Here's a story illustrating how people think Clear Channel harms consumers.

Saturday, March 30, 2002

Big media consolidation does not matter This sincere diatribe against the evils of media consolidation reflects how the "Left" can often be it's own worst enemy. The argument is that single ownership of multiple distribution networks and content will lead to less diversity in media. Apart from the empirical fact that diverse content has been exploding in every channel since the 70s, there are also simple economic truths behind what's going on and why the end result will not be bad.

Firstly, there are economies of scope in owning multiple channels because then you can re-purpose content. The only reason this hasn't happened already is the certificate of convenience and necessity regime cooked up by the government and radio monopolies as they made their migration to TV after World War II. Deregulation allows these economies and will result in more, cheaper content.

Secondly, there's no reason a vertically integrated content producer and distributor should exclude rival content or refuse to license it's own content to other distributors. Once you factor in opportunity costs it really makes no difference one way or the other (unless the distributors do some important local marketing, in which case consumers benefit from that integration). Media companies are not out to control the world, they just want to make as much money as possible.

And thirdly, these soft-headed whines ignore how the power shifts from one part of an industry to another. ClearChannel is a popular bete noir because it toasts small local stations. But small local stations are in a poor bargaining position against the recording industry cartel, and ClearChannel can do a much better job of negotiating better terms from them. This helps consumers.

Most amusingly, the authors don't recognize how Fritz Hollings' (D-South Carolina and wholly owned subsidiary of Disney, Inc) actions against FCC deregulation protect the very media cartels they excoriate. FCC regulations maintain the high federally mandated barriers to entry the incumbents currently enjoy, and deregulation would lower them. (via Rebecca Blood)

Friday, March 29, 2002

KaZaa A Netherland judge overturned an earlier decision arguing that since KaZaa offered legitimate uses as well as unauthorized file sharing and had no central server, it was not liable for its users' illegal acts. But the lawyers don't seem so worried.

More of a concern to them is Judge Patel OK'ing antitrust investigations into the music cartel. Copyright, like any monopoly, harms society by creating dead weight loss. One way to reduce this, short of dismantling the monopoly, is to price discriminate, which reduces efficiency but also transfers consumer surplus to producers. The problem is that consumers seek surplus, so if you take it all away, they'll substitute something that gives them more. So merely transfering surplus is not enough, you also have to collude to eliminate all substitutes. And this is illegal under antitrust, but we all know what that's worth these days (the SSSCA also enforces collusion among hardware manufacturers, no one company could survive crippling its products alone).

Does DeCSS circumvent trade secrets? This article makes little sense, but I think its says the DVD cartel is arguing that DeCSS code should not be published because it's a trade secret. My guess is that a DMCA based ruling was overturned on First Ammendment grounds, so lawyers are going the trade secret route. I think the judge will say "if it appears on a t-shirt, it's no longer a trade secret."

Microsoft does not know how to sell into enterprise Microsoft (and Uniysis) are going to run ads telling people not to use Unix because it locks them into proprietary, expensive systems. Apart from the outright gall, you gotta wonder who these ads are going to effect. Enterprise software is not soap, and I fear for any company that makes its purchasing decisions based on 30 second shills that interrupt Seinfeld reruns.

When I interviewed at Microsoft (yes, yes) most of their questions revolved around selling shampoo or fizzy drinks. But the skills required to push sugar water are entirely unrelated to selling enterprise software into enterprise (which the company wants to do, but clearly doesn't know how). I wonder if Microsoft's marketing department knows how irrelevant it's been to the company's success. The company's strategy department, on the other hand, knows exactly what it's doing.

Thursday, March 28, 2002

Uniform and non-discriminatory terms People beleive that uniform and non-discriminatory terms are good things and a suitable means to control monopoly power. They are wrong. Vital monopolies like Microsoft can use them to force unfavorable terms on everybody (first Sony with patents, now Dell with returns). Also, if monopolies are allowed to price discriminate, they can charge lower prices to the more elastic section of the demand curve. This transfers wealth from consumer to producer, but it's also more efficient, so probably a good (albeit distasteful) thing overall. Finally, non-discriminatory terms are a way cartels can enforce collusion through contracts, even though such contracts are illegal. Someone slipped this language into the Telecoms Act of 1996 which protected the ILECs (ironically, clueless CLECs may have fought for it's inclusion).

Wednesday, March 27, 2002

Eisner Whether or not Eisner is worth 11% of Disney's profits is between the company and its shareholders. But he treads on dangerous ground in his FT op-ed against "piracy." While Lincoln may have approved of copyright under it's original 14 year duration (with an optional 14 year extension) he would have taken a dimmer view of the current life+70 year monster.

Article 1 section 8 of the Constitution empowers Congress "To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries." It's pretty clear that retroactive extension and essentially unlimited terms do neither, no matter how loudly the content hoarding industries (Hollywood the RIAA cartel) argue otherwise. In fact, if you're going to get all misty eyed and start basing ideas on what's best for society, it's worth noting you the current copyright regime costs $57 Billion and that's just for CDs.

Finally, as Dave Winer notes, there's also the cost of enforcement. The CBDTPA nee SSSCA would essentially burden the technology industry with the cost of policing unauthorized digital copying. This cost is ruinously high, as content owners well know -- note that no individual has actually been charged with copyright violation, only tools have been attacked via the DMCA. If the cost of policing a law is greater than the harm breaking that law does, then it shouldn't be a law at all.

Fundamentally, Eisner is wrong. Ideas aren't property. Copyright owners are free to persecute unauthorized copyers under exiting laws, and should do so (isntead of outlawing tools). This is not arguing for no copyright, just a lower copyright tax. Copyright, in it's current bloated form, harms society and should not be supported.

Tuesday, March 26, 2002

Napster for academics My entry in a Chicago business plan competition was software that helped academics share their research with each other. Academic journals make obscene profits (libraries are very price insensitive) and I wanted to shift some of that surplus to consumers (the scientists -- who also happened to produce the content). The judges were not impressed by my idea, but it lives on as George Soros's Open Access Initiative.

The whining you hear in the article is the handful of academic journal publishers moaning about their profits (maybe) going down. But they shouldn't worry -- refereed journals bring prestige because they are hard to access, and thus are valuable to academics.

Monday, March 25, 2002

MSFT+DRM As I (and others) pointed out ages ago, Microsoft's goal to own all bits means it will build DRM APIs into it's OS. The current settlement requires Microsoft to disclose middleware APIs unless they relate to (among other things) DRM. So I wouldn't beleive Allchin's claims that Microsoft can no longer build in secret DRM APIs to establish Windows Media Player as the standard and kill Real.

Underage Darwin Here's a nice post about how hard it is for anyone to contribute to Darwin, Apple's open source core, particularly if they're under 18. But the *BSD community has always been pretty closed, even though the code is open. A single company must take ownership of all patches if it wants to dual-license the code, which is what Sleepycat does. GNU/Linux has so many contributors, that they probably couldn't dual license even if they wanted to. The under-18 non-disclosure law once upon a time might have protected minors, but in this day and age is just a barrier to entry for young coders.

Friday, March 22, 2002

Patents vs Copyrights The costs of patents are (kind of) clear, but not copyright. So I hope Maxygen (uses irritating dhtml) is successful at copyrighting gene sequences by turning them into songs. It's a sleazy use of a sleazy law and makes the costs of interminible copyright more public.

Broadband There's a truly idiot article in US News claiming the current bandwidth glut will become an acute shortage, which ended up on Slashdot. It gets both the technology and economics wrong. All the sunk fiber will be lit once it's economically feasible to light it. Lighting fiber has higher fixed cost per pair than sinking it, but lower total capital outlay.

Scientology vs. Google I'm glad the Church of Scientology used the DMCA to block Google archiving some site in Norway that talks about its (copyrighted) space alien nonsense. It highlights nicely why the law is so corrupt.

Thursday, March 21, 2002

What's good for the Internet? Doc asked if Tauzen-Dingell was good for the Internet. I think that innovation and efficiency are good for the Internet, and as Tauzen-Dingell promotes both, I'd argue it's also good for everyone's favorite network-of-networks.

What's good for competitors is not always good for customers. For example, if there is some minimum efficient scale of production, marginal competitors would like to keep dominant players down, but end consumers might prefer the lower prices offered by an oligopoly over high-priced choice. Or they might not.

The local loop is a natural monopoly. TA96 made incumbent Bell monopolists offer access to the local loop for marginal cost plus a little (read the gory details here). CLECs (competitive local exchange carriers) took advantage of this by offering end consumers stuff more cheaply. Note that they did not invest in upgrading equipment, they just freeloaded of the Bell's sunk costs, AND they picked the best customers ("cream skimming") while telcos had to serve everybody (through Universal Access provisions). Some champion the CLECs as being heroes of the people, but all they did was remove the incumbent's incentive to upgrade (which it had, even as a monopolist). CLECs themselves never had any incentive to invest in equipment, so sensibly never did. Removing the incentive to invest and upgrade is bad for innovation, and therefore bad for the Internet.

A more arcane point is that CLECs may have been less efficient than the incumbent monopolists. As the Bells priced at monopoly price (or even average variable cost), CLECs could undercut them by pricing at *their* marginal cost, even if that was higher than the incumbent's marginal cost. A system that funds inefficient entry is wasting money, and thus, inefficient. This is bad for innovation, and therefore bad for the Internet. But this point is boring, let's move to something jucier.

Taxes are bad for efficiency (create dead weight loss), and therefore bad for efficiency. "Universal Access" is essentially a tax on long distance that subsidizes local calls. It's grown over the years and currently burns 40 cents for every dollar it raises (incumbent monopoly pricing burns 11 cents for every dollar it raises, income tax burns about 25 cents for every dollar it raises). This is clearly bad, and is one of the reasons CLECs found cream-skimming to be profitable. TA96 called for Universal Service of broadband to be extended to schools. I'm not sure whether Doc thinks this is good for the Internet or not. It's certainly wasteful, but it's important kids have Internet access (and it's worked pretty well, particularly in the poorest schools). This is the sort of decision society needs to make for itself.

What is absolutely bad for the Internet is vetoes on innovation. Any kind of a tax on broadband provision eliminates marginal areas from receiving it, costing the entire consumer surplus of those populations (which will be huge). Digital television has vetoed non-digital TV use of that spectrum, eliminating services we haven't even dreamt of yet. The satellite industry has paid Congress for a clause stipulating that their spectrum will never be allocated by auctions, vetoing all other uses for that slice of real estate (and robbing the country of billions in auction fees). All of these are bad for the Internet, and innovation. Tauzen-Dingell does none of these things. And while some overheated commentary argues telcos will just block competitors from using their networks, they're just wrong. Telcos will simply charge competitors the monopoly price, forcing efficient entry, while reducing incentive to innovate by less than alternative pricing schemes.

Tauzen-Dingell removes the quid pro nihilus regulations TA96 offered CLECs, and lets local incumbent monopolies act like rational local incumbent monopolies. Is this ideal? No. But it's better than making them act like local incumbent monopolies with no incentive to upgrade equipment.

And how is the Internet doing through all of this? 90% broadband growth last year. Fairly stiff competition between the equally hated cable company and telco. CLECs are toast (but they did no infrastructure investment anyway). All in the face of fairly flat dial-up growth. I don't think this is bad, but Bruce Kushnick would disagree. But then again, he thinks there isn't enough fibre optic in the ground (even though 98% of what exists is unlit), doesn't see cable modems as a DSL competitor, and thinks CLECs should be supported, no matter how little they do to improve the infrastructure the Internet runs on. (He's also partially funded by the Morino Institute, that I once did a little work for.)
Link to this column

Doc implies I'm right wing Maybe I was unfair to label Doc a "lefty". He describes himself as an anti-war, registered independent, libertarian. He also beleives there is an Economics of Altruism and is currently adrenalized about protecting Internet freedoms.

Regular winterspeak readers would recognize protecting Internet freedoms is also something I'm adrenalized about. But not having Doc's experience in the field I must rely on Chicago school Law & Economics to figure out what's "good for soceity" (Doc assumes if it's good for the Internet, it's good for society, and I mostly agree). So to answer Doc's question, "Is Tauzin-Dingell good for the Internet?" I'd have to answer "Yes." It prevents inefficient entry (which TA96 encouraged) and gives telcos the resources to upgrade equipment. (I'll write up broadband/telco regulation later today.)

Now I don't like telcos, and I don't trust telcos, and I'd keep my eye on them, but TA96 was bad, unworkable legislation, and we should be glad if it goes. (Btw. If anyone cares I'm about as right wing as The Economist.)

Wednesday, March 20, 2002

Wrong about telco reg Lefties like Doc often evaluate laws by how it moves their moral sensibilities. Here are some of Doc's headlines that sum up Tauzin-Dingell regulation: it stinks!

But you should judge regulation on the effect it has on people. While I'm no fan of telcos, I won't block good regulation because it benefits companies I don't like. In essence, Tauzin-Dingell undoes 1996 pricing regulation (TA96) that mandated local loop incumbents 1) provide access to bottlenecks at just and non-discriminatory prices, 2) offer unbundled network elements (UNE) for resale and 3) allow local carriers to enter long-distance if they open their local loop. Sounds good, huh? Let's see how it falls apart.

1) Provide access to bottlenecks at just and non-discriminatory prices
The problem is that no one can agree to what such a price should be. The local loops are natural monopolies, so they push for the efficient component pricing rule which basically leaves them free to exercise their natural monopoly pricing power. Although this sounds evil, it encourages efficient entry and innovation, and will push down prices in the long run (albeit not quickly). But at least the incumbent has no inventive to block new technology, which in the grubby telco world, is quite novel.

The law passed in 1996 uses TELRIC pricing instead, which tries to set the cost somewhere between monopoly price and marginal cost, which allows the incumbent to theoretically recoup the cost of upgrading infrastructure. Unfortunately, this encourages inefficient entry and because it's "forward looking" (i.e. based on the marginal cost of current, best of breed services) reduces the incentive to upgrade equipment (because in practise incumbents can't gaurantee a reasonably return on their investment). Unsurprisingly, equipment upgrades have ground to a halt and the FCC has a huge line of arbitration cases filed by entrants saying the incumbent screwed them through unfair pricing.

2) Offer unbundled network elements (UNE) for resale
Heh heh heh--there's a clause in this section stipulating the incumbent has to offer UNEs at the same rate to all customers. A monopoly can only earn one monopoly profit, so it wants to restrict the quantity to make that profit. But if it can price discriminate, it can actually produce more AND make more money than if it was just a monopoly. Of course, it does this by reducing the amount competitive local exchange carriers can make but as a consumer I have no interest in subsidizing those guys, I just want cheaper rates. This non-discriminatory clause, while it *sounds* helpful, actually keeps the incumbent from cheating and so results in it restricting quantity, resulting in higher prices for end consumers, which is bad. Ahh, legally enforced collusion at work.

3) Allow local carriers to enter long-distance if they open their local loop.
This is a carrot, that says if you do 1) and 2) you can have 3). But long distance isn't a natural monopoly the way the local loop is, and the profits there are pretty slim, so it's not clear why this should entice local incumbents. Moreover, if there was significant monopoly power in long distance, joining the local and long distance parts of the service would remove the double-marginalization effect you get when two monopolies are stacked on top of each other. This would reduce the price I pay as an end consumer, which is good. So why aren't they allowing this anyway?

If economics doesn't move you, how about what you see with your owns eyes? In its six years of legislative existance, TA96 has done nothing to support competitive entrants into telco markets, created unbelievably long lines for FCC arbitration, and broadband is growing at a healthy 90% a year anyway. TA96 was bad regulation that hurt consumers. Be glad it's gone.
Link to this column

Tuesday, March 19, 2002

Irony In a recent post, I mocked Chernin of News Corp for making particularly misleading comment about broadcasting trade secrets. My friend CD points out that a Newscorp subsidiary in France is under fire for broadcasting stolen Canal Plus code (NYTimes, subs required).

Monday, March 18, 2002

Not the end of the world Nice article here by Andrew Bowie pointing out that the music industry is not collapsing, and the movie biz is positively radiant. Journalists like claiming the music industry is vanishing because it makes good copy, and the recording cartel harps on about so they can push through draconian legislation. In reality, an artistic dry spell after four years of bumper crops means CD sales are down, while online movie copying is so horribly hardly anyone bothers (and if the cartel released DVDs along with movies, even fewer would).

Similarly, broadband is another phenomenon industry pundits (including yours truly) pooh pooh as growing at a snails pace. But in 2001, broadband grew 90% while dial-up growth was virtually stagnant. So when any company claims broadband is not happening, look for legislative intent. If any pundit claims it, look for cluelessness. But please note "broadband" is defined as at least 300kbps both ways, which is about an order of magnitude lower than what movies require across the internet. So no video on demand or "convergence", just broadband (which is the same as the dial-up internet except less irritating).

Saturday, March 16, 2002

Just for fun I think this is just hilarious.

Friday, March 15, 2002

Software gets pirated too Tim O'Reilly picked up on the quote I found most obnoxious in the recent Times piece on Holling's SSSCA legislation. Chernin, of News Corp, bleated
"Let's say I decide to broadcast on my network the code for how to make Intel chips or Microsoft software... I think they'd find a way to stop it."
Ignoring the fact that a simple lawsuit would stop this (instead of banning TV), the point is that software makers have their products pirated all the time and went down this anti-piracy path in the 80s. Code, after all, is as digital as mp3s (if not moreso). Anyway, Tim pick's up on this and nails it:
Consumer behavior gave marketplace advantage to companies that didn't use copy protection, and after a relatively short time, the industry got over its fears and got back to offering products that people were glad to pay for.
Go Tim! (link from Archipeligo)

Thursday, March 14, 2002

Spectrum I was flipping through an old Release 1.0 and read Kevin Warbach arguing that spectrum should be an open common. He claims that you no longer need to worry about interference because new technology means node devices are much better at sifting out signal from noise, so why not let people broadcast at any frequency?

Currently, spectrums is licensed through government auctions. In the bad old days, it was allocated through a beauty contest, so graft and waste was rampant. Auctions have been spectacularly successful except for NextWave, but that's an anomaly because the rules were changed for that bidding round (so anyone who points to NextWave to critique auctions is being lazy). Telcos moaning about high auctions prices are just upset they had to pay the fair market price, and should be ignored.

Moreover, spectrum owners should be free to resell spectrum to whomever they choose for whichever purpose, which would result in inefficient spectrum use being sold to efficient spectrum users. In the US, broadcasters have their spectrum allocated for their use and their use only, so they can't get rid of it even if they could do so profitably. Satellite providers bribed Congress to write into law that their spectrum would only be allocated through beauty contests (so bring on the pork!) The most productive and efficient user of spectrum will be willing to pay the most for it, and so inefficient owners have economic incentive to sell it to them. Beauty contents and pre-allocated use destroys this efficiency.

But back to Warbach. It's true that more power (and money) in the nodes lets you get more juice out of spectrum, but that's not the same thing as saying there is no interference. Also, he argues the rules to stop cheating should be coded into the hardware, but this is easy to crack and break. Spectrum allocation started because some religious radio station in Montanna was drowning out everyone else around--nothing but software stops this happening in Warbach's world. More auctions, an aftermarket for spectrum, and non-prescriptive property rights would improve spectrum utilization more than hippie WiFi networks. But hippie WiFi networks are probably better than much spectrum use now.

Wednesday, March 13, 2002

SSSCA and the New York Times Jonathon Zittrain is an assistant professor at Harvard Law School and was roommates with one of my professors here at Chicago, Austan Goolsbee. Zittrain has a NYTimes Op-Ed piece where he talks about the SSSCA, and how getting computers to do less would make them more reliable.

While it's true that complexity is the enemy of reliability, and one way to reduce complexity is to strip out functionality, it is also possible to have secure, flexible, powerful systems (see Unix). And while I'm sure there's a market for simpler Internet appliances that do less, Zittrain fails to point out is that the SSSCA aims to outlaw the general purpose computer entirely (along with general purpose operating systems, like GNU/Linux). And we know there's a large market for general purpose computers.

If people want to buy an uncomputer (like a TiVO) they're welcome to do so. But outlawing the general purpose PC is insane.

Copyright cartel insanity Salon has two articles which are nice to read back to back: one is how the recording industry cartel is pushing for complete control over all distribution to stop people listening to music for free, the other is how that same cartel is moaning about how expensive it's become to bribe radio stations to play music (so people can listen to it for free).

Tuesday, March 12, 2002

Will GNU/Linux disrupt Sun? They key to whether a potential disruptive technology is a threat has nothing to do with how good the incumbent's technology is. It has to do with whether the disruptive technology is improving faster than the market can absorb improvements. The incumbent could have better technology and innovate more rapidly and still be disrupted by a worse, less innovative disruptor so long as it eventually catches up with the customer. Is GNU/Linux disrupting Sun on the workstation? Maybe. How about Microsoft on the desktop? How much has the Windows desktop really improved for the end user since '95? How much have KDE or GNOME improved over the last two years?

Comingling vs Bundling vs Tying It's ironic that if, back in '95, Microsoft simply offered two versions of Windows, priced identically, one with I.E. and one without, the antitrust trial would not have happened. Doing this was not a big deal for Microsoft, OEMs would have usually included I.E. (especially is arm twisted a little) and maybe included Netscape as well. But they didn't and the rest is becoming history as we speak. If they had only bundled the two products together instead of tying them, the court would not have had a case.

Some of Jim Allchin's testimony outlines how pointed litigators are getting about "componentizing" code. Clearly someone who knows about technology is designing this line of questioning. At stake is how much power OEMs have in piecing together OSes for their customers. If Microsoft "componentized" Windows, customers could sub in best of breed middleware, or just use simpler, cheaper middleware options, and probably be happier (they would certainly be more secure). Doing this would lower the application barrier to entry that protects their Window's OS monopoly, so they are against it, denying that it's even possible. This claim is a lie, and if not, should worry users. Monolithic complicated systems are insecure, unstable, and bad network citizens.

So, bringing events back to 1995 -- the court may ask Microsoft to offer a componentized version of Windows.

Good Economics This article is so smart I thought I'd include it, even though it has nothing to do with computers. When I lived in New York, it impossible to catch a cab around midtown at 4pm. The city regulates taxis by limiting the number on the road and by setting what rates they can charge. As my "no taxi" problem illustrates, the city has too few cabs. Fools are suggesting that NYC increases the fares cabs can charge, but this will result in fewer cabs (they need to work less to get the same income) and more expensive cabs (price sensitive customers use subways and buses anyway) not more cabs, which is the problem. The effective thing to do would be to increase the number of cabs by issuing more licences, which currently auction for $200,000 a pop and would raise needed revenue for the city. But the fleet owners who monopolize this resource want to keep the quantity restricted, so will lobby against any increase. New Yorkers should expect a needless increase in fares along with an small license auction. They should not expect available cabs.

Monday, March 11, 2002

Andreeson grows up In this Economist article about the new, grown up Marc Andreeson, the Internet poster child says
The trouble with selling software is that vendors end up in an adversarial relationship with their customers. The customers know that the vendor needs to meet its quarterly sales target, and is often dependent on a few large sales to do so. The result is a frenzy of last-minute negotiation and price-cutting at the end of each quarter
He gets it quite wrong. Any public firm trying to meet quarterly analyst targets has the above problem. Big companies like GE dip into secret stashes so they can hit earnings to the penny. Tech stocks' value comes from revenue growth more than profits, so they need to show rapidly increasing shipments. Without big cash reserves to dip into this means they stuff the channel at the end of every quarter and give big discounts to the customers so they can make earnings.

The adversarial relationship comes afterwards once the buyer is locked into a single vendor. The software company can now squeeze money out of the buyer at will through pointless new features, draconian licensing laws, and forced upgrades (a la Microsoft or Oracle). While the buyer might have forced enough concessions by getting an initial deep discount to make the eventual contract value neutral (i.e. no one makes money on it), it's not a good environment to create quality software.

Saturday, March 09, 2002

Apple in India Apple's future depends on OS X, and since people don't buy operating systems, they buy applications, that means OS X needs lots and lots of apps. But hardly anyone owns a Mac, and many Mac users are still quite happy with Mac OS 9, 8, 7 etc. And since no one has OS X, why should anyone write programs for it?

Gasee didn't understand how tricky this problem is, but Jobs does. That's why he included the Classic layer to help OS 9 users move up to X. Apple also leaned on key software providers to launch OS X apps in anticipation of the market (Microsoft, Adobe). Darwin means OS X users can tap into all POSIX compliant code. And finally, as Apple is a software company, they released the iSuite for OS X. Now, they're hiring cheap Indian programmers to port applications to OS X. I don't know what they focusing on--Star Office? Games? the GIMP? Vertical apps like BLAST? If people have any idea, please write me at

Friday, March 08, 2002

Hardware vs. Software The peculiar thing about OS X is that it highlight's Apple's UI engineering skill--they've put a fantastic GUI on Unix (here's a fun history on operating systems) but have structured the company around end-to-end integration. According to various analysts, OS X's adoption will be stymied by it's reliance on Mac hardware--in particular, Moto's PPC.

Is this true? Quite probably. If you're not a Mac user, switching to the platform requires an investment in Apple hardware which, while reasonably priced, is still expensive. But Apple wants people to buy their hardware, so they'll pass on this opportunity. The real value in owning hardware, OS, and applications is appropriate investment in smooth integration across all three, which users of USB/Firewire peripherals, iMovie, iDVD, iTunes, or iPhoto can attest to. But is the value Apple creates through end-to-end integration larger than what it could capture by trying to displace Windows on x86? At first blush, the answer seems a clear no--displacing Windows is a much bigger deal than making it easier to shoot home videos. But MSFT would probably pull support for Office OS X which, to be honest, would toast the company. (Thanks to Mark for the link)

Thursday, March 07, 2002

eMusic economics It's worth reading through this NY Times puff piece on the music cartel's online offerings. First of all, note how you need Microsoft products to run anything at all (Windows, I.E., Windows Media Player). I wouldn't be surprised if it's being served of some flavor of NT, or will be soon. This is a glimpse of the Microsoft future where they control every bit from creation to transport to distribution.

Secondly, it's worth looking at what a bad deal these services are. The marginal cost of producing an album is the same as producing a single. But willingness to pay for singles is lower ($4) than albums ($15), so the profitable thing to do is phase out singles for albums--which is what has happened. But albums are actually a terrible deal--most of the songs on them are lousy. But through this sort of bundling, consumers get more value overall (higher willingness to pay) thus distributors can charge a higher price (which, keeping cost constant = higher profit).

A fairly priced single, at $1.50 a song, is stiff competition to the album, so note how painful it is to put together partial albums in these services. Also, the $1.50 single ($15 album with ten tracks) lets you hear these songs in perpetuity -- at a 5% discount rate this works out to 7.5 cents a year, or 0.625 cents a month (I think, sloppy maths here I'm afraid, it's late). This clearly is much less than the extortionate amounts being charged per month for music now. These online services are a joke. Any real competition would gut them instantly.

Eldred vs. Ashcroft Here's a really nice article examining the Supreme Court review of the Sonny Bono Copyright Extension act. The upshot: the court will probably rule retroactive extensions are unconstitutional (quid pro nihilo) but very long terms probably are not--they're just bad policy. It's difficult to think of a moneyed interest group willing to fight for the public domain.

Wednesday, March 06, 2002

Patent Madness The French seemsto be against Europe's mad US style software patent initiative, arguing that it's bad for innovation in general, and free software in particular. It's interesting that a government singles out free software as an important factor in considering legislation.

The problem with patents is that their effect on innovation is unclear, as they both increase the value of innovating and the cost of inputs of innovation. Moreover, there are strong incentives to innovate software and business methods anyway, so they fail the acid test for intellectual property.

Finally, it's worth noting that getting a patent is quite different from defending a patent. The Patent Office will readily hand patents out for any old thing, but the "non obvious, useful, and novel" standards are called into question when defending patent claims in court. This is inefficient because it forces people to litigate against patents to demonstrate their invalidity.

Tuesday, March 05, 2002

Compulsory licensing Ages ago I wrote about how Napster could work, essentially by a flat-fee subscription that's distributed according to usage. The business side of this is compulsory licensing, which the linked article writes up well. This structure means that artists can be compensated without record company interference, and will therefore be blocked by the RIAA. However, given the extreme complexities in licensing (as the RIAA themselves are discovering in the Napster case) it might be legislated anyway.

This sort of licensing permits copyright holders to be compensated for their work without giving them a veto over new technology. This is incredibly good. How it's implemented, though, is crucial for how much good it does overall. Some people talk casually about taxing ISPs (especially broadband), ignoring the very high fixed costs of entry, and therefore not considering how marginal markets will be unserved if service is taxed. The loss to society in this case is the entire consumer surplus those people would get. So the most efficient tax would be a small one on an existing good with highly inelastic demand that is currently lightly taxed. So not cell phones or long distance.

Monday, March 04, 2002

Apple gets out of the printer system business First the kernal, now the printer system business. It's nice to see a company stop making software it isn't much good at and focus on areas where they can add distinct competitive advantage.

Friday, March 01, 2002

Hollywood vs the World The unholy triumverate of Valenti, Eisner, and Hollings, set the stage of reviving the draconian SSSCA, legislation that would outlaw fair use, computers, the internet, and open source. Intel's Leslie L. Vadasz put it best in saying "content, once captured in "unprotected" form, can never be put back in the
"bottle" and protected against copying on the Internet."

The key to the SSSCA is that it outlaws the public domain. Companies are free to encrypt their content to their hearts' degree, but this degrades the experience of using it compared to using open content (either under copyleft or public domain). In a competitive market, open content will win.

There's already been a backlash against crippled CDs, and there will be a similar backlash against crippled hardware. The central point is that draconian copyright reduces the willingness to pay of consumers, limited the price companies can charge for the content, and by extension, their profits and revenue. Content's quantity is already very restricted, so this will mainly impact on the utility of PCs. Of all the technology companies, only Microsoft might be interested in this legislation since it outlaws GNU/Linux, the only threat to its OS monopoly.