Tuesday, July 30, 2002

Snigger Thomas Krul complains about how the Linux GUI is not "pleasurable" enough and advocates an open-source GUI community to complement its developer community. I'm all for open-source GUI work, so long as Thomas Krul is not on it. His post is split into two columns, fine for a magazine, but terrible for a webpage because it forces the user to read all the way to the bottom, and then scroll all the way back to the top, and then read all the way back to the bottom again. If this guy can't get a page of text right, or worse, thinks pointless scrolling is "pleasurable", I don't want him anywhere near my GUI.

Monday, July 29, 2002

Hollings is at it again Once again, congressional adjunts to the entertainment industry are trying to outlaw computers and revoke the Internet. One bill authorizes denial of service attacks on servers thought to be trafficking in copyrighted material, while the other makes adding DRM to non-DRM content illegal. Hollywood and the music industry know that free content will beat crippled content in any competitive market, which is why it is trying to outlaw fair use and the public domain. (via Tomolak)

Friday, July 26, 2002

One more thing John Patrick has a book out called Net Attitude. And there's an archived tnbt interview with Jeffry Rayport, who was once upon a time my boss. Small world.

Thursday, July 25, 2002

Patent Madness I went through a "patent scrub session" today where basically we spoke to a lawyer about our software work, and I guess he's going to go and file patents on everything. Sleazy business. But given that patents can be filed, companies like IBM need to file all they can.

IBM recognizes that patents hamper innovative software development, which is bad for the industry because little startups (which can be bought) are much better at coming up with new ideas than big companies. No new ideas, no new products to sell. Moreover, IBM's patent portfolio means nothing against other big guys, because they have portfolios too and it all just gets cross-licensed. And it means nothing against little guys, because what fool startup would dare file a patent suit against IBM?

Answer: a burnt out shell that has no assets except patents and is essentially no longer a company, but a bunch of lawyers looking for settlement money. This whole software patent thing is a ridiculous arms-race that distracts the industry from the work ahead of it.

Wednesday, July 24, 2002

Lunch with John Patrick John Patrick was one of the guys who got IBM turned onto the web, and he remains unfashionably pro-Internet today. I had lunch with him today, and then heard him speak to a larger group about where he thought the Internet was going.

John's a pretty famous guy, and was instrumental in getting IBMs Fortune 500 customers (which is all of them) "e"-iffying their businesses. To this day, he feels the Internet holds tremendous promise for business and that the digital transformation is only "2% done". He points out lame corporate sites forcing visitors to dial some 800 number, M-F 9-5, as indicating how corporate america still doesn't get it.

Corporate America may not get the Internet, but I don't think it needs to. Offering better service over the Internet for an established company is just a form of non-price competition, which is fine if you're the only guy doing it but kind of stupid if all your competitors are just going to imitate you. And 24 hour call centers are very imitable.

I think companies are right to do as little as possible in moving their business online. They'll have to get there eventually, but rushing is too much like competition, which is bad for everyone's profits. The executive clueless that John derides is enforcing industry-wide non-compete agreements, which is pretty funny.

Startups built around the network are the only things that will change mainstream business. eBay is shaking up retail, durable goods, and newspapers (who make most of their money off classifieds); the Napster successors are shaking up music, weblogs are shaking up trade magazines (but not newspapers, the Op-Ed page doesn't make any money), and GNU/Linux is shaking up operating system companies. These are the true Internet organizations where new value is being created, old line companies becoming moving to the Internet are just passing it onto their customers.

Tuesday, July 23, 2002

Isenberg's "SMART" letter is dumb Rob Isenberg is some telco industry pundit who runs an industry newsletter called "SMART" which is pretty funny, because it's pretty dumb. His latest edition rambles on about bad debt, overcapacity, and a change in telco business model.

First of all he makes the mistake of thinking that financing a five year investment with 30 year bonds is somehow a problem. It's not, the finance structure has zero effect on whether or not the equipment investment was profitable. I will repeat: the duration of financing for equipment is unrelated to its expected longevity, and the decision to invest in equipment is unrelated to the longevity of its financing.

Then Isenberg goes on about why telco has too much overcapacity, apparently without realizing that this is the real reason investment in infrastructure has dried up. If he engaged brain, it would be clear that investing in more when there's already too much is stupid, but instead he talks about "glut" and "profit" as if the two were unlinked. Entry is only sensible when there is scarcity allowing the industry to charge supra-competitive prices, otherwise investor money should go elsewhere.

And finally, Isenberg means goes on about telco business model. Before deregulation, the government set rates, business subsidized commercial service, and urban subsidized rural users, but things became unclear because deregulators could not (and still cannot) decide whether they want competition at the infrastructure level or services level (Isenberg illustratively wants both and neither). This regulatory environment determines business model, because with telco you can either get competition at the service level or the equipment level, not both. The telecom act of 96 set competition at the service level by letting entrants pay marginal cost for infrastructure access, and people like Isenberg complained that phone companies are trying to add services and weren't investing enough in equipment.Tauzen-Dingell legislation tries to undo TA96 and shift competition back to the infrastructure level, but people like Isenberg then complain incumbents block entrants by demanding outrageous access charges.

Telco executives in years past have wanted to avoid becoming "dumb pipes" but I suspect this is because the hanker after glamour only content business can provide. But shareholders care about profits (just ask Time Warner - AOL) and ultimately a nice local dumb pipe monopoly can generate those with no problem.

Swag Just got a most excellent Peace, Love, Linux shirt from IBM. Nice!

Monday, July 22, 2002

Real opensources client Real Networks is a media company that sells eyeballs to advertisers, which is why its client is so user-hostile. Microsoft is an operating system company that's extending its monopoly into applications through middleware tying, which is why its clients are so standards-hostile. An OS monopoly will clean annoy-ware's clock any day, so Real seems to be opening up its client code.

Since the client is free-as-in-beer anyway, I don't see how making it open as well will help their position, as making it free-as-in-speech just seems to open the door to a non-irritating client. This would kill their (no doubt enfeebled) advertising market, leaving them to sell servers. And who knows, the media business was always a non-starter so maybe now applications will choose to incorporate a Real client, boosting server sales. We'll know when the licensing details are released.

Friday, July 19, 2002

Lunch with Andy Had a great lunch with O'Reilly's Andy Oram today. Mostly we discussed the emerging Internet Operating System (I heard it first from Tim) and what that means for P2P, collaboration, and security. It's more than I can do justice to in a short blog, but I'll write something long on it soon.

Oh yes! I was shocked to see this post on Good Eats' Alton Brown on Slashdot. One of the reasons I bought TiVo was so I would never miss his show, I recently bought his book, and yes, I keep bricks in the oven. I'm pleasantly surprised that this interest is normal and the reference group turns out to be slashdotters.

Thursday, July 18, 2002

Bye Bye Northern Light Had lunch today with a friend who worked at Northern Light. The CEO was into race cars, so he had his company sponsor the Indie 500 for $5M. Where were the shareholders or the board? The company's been reduced to a skeleton technical staff and were bought by Flip Flipowski, a Chicago technology guy who I think is bogus.

More Macworld I saw the rest of Jobs' keynote today, and I was impressed by how carefully they've been thinking about what people are using their tools for.

Something I've noticed at Lotus is how engineers start by trying to solve a problem, but then notice that the problem is just a particular instance of a larger class of problem, and then they start to try and solve that, until they realize they can abstract that problem to something even broader. By this time any useful application is long gone and they end up producing some piece of infrastructure -- fine to build other stuff on but useless to end users. Which is how Notes feels to me -- ugly scaffolding.

But Apple focused on the applications. There's a reason iChat and iTunes links into rendevous--it makes rendevous useful. Yes they have a database built into the OS, but it handles people, and today I mostly use my computer to share stuff with others. And Sherlock's web services plug directly into every day applications like restaurants and movies.

I actually saw the product Daniel refers too in an Apple store and didn't know what it was, so I looked it up online and discovered it's a $45 volume knob. That clicks. I want one, but only because it's so shiny.

Wednesday, July 17, 2002

Macworld NY Stopped by the Mac store in the Cambridgeside Galleria this morning to see what His Steveness had to say about Macs. Three major things: OS X.2, .Mac, no G5s.

I'm glad Apple decided to focus on the very real performance issues with OS X. All the eye-candy slows the GUI down and results in a performance inferior to OS 9 (at least on my Pismo) and is reason enough not to switch. Jaguar (Jag-wire) seems to fix that, and I'll probably buy new hardware once it comes out. Moreovoer, Rendevous strikes me as being one of those killer technologies that change the way we think about computing, and I am *very* excited about having it on my machine.

Apple has also decided to charge $100/year for the previously free iTools services. I think that charging for these services is correct, but the pricing is wrong. Unless you make more money on a product than it costs you to offer it, you probably shouldn't be offering it, so if customers find iTools useful they should be willing to pay for it as .Mac. Moreover, customers paying for services directly means that the vendor has to care about the customer experience, which is not a bad thing in this world of crappy software. But Apple should have been more flexible than $100/year or else -- it could have garnered more subscribers than it will.

Finally, although Apple upgraded the iMac, the G4 tower remains the same. These two machines are so close that too many customers are trading down is too attractive. Reducing the power of the iMac is out, which means Apple needs to raise the bar on its high-end machine. That did not happen this expo.

I'd also like to mention how neat I think Apple's consumer facing web services are in Sherlock 3. Web services are infrastructure, and no one buys infrastructure, so it's wonderful to see someone (Apple) thinking hard about applications that make life better for end users to drive web services adoption.

Tuesday, July 16, 2002

Old projects never die Here's a great article tracking how a secure playback mechanism for DVDs morphed in Palladium, Microsoft's patented Digital Restrictions Management operating system. From my time at IBM I can see how projects drift out of research, and then in and out of products. But one part is clear: Palladium does nothing to improve user security and everything to protect media from being copied. It's not clear if unsigned media can be played though.

Monday, July 15, 2002

Office for Mac OS X Microsoft's threatened to cut support for Office on Apple because of dissappointingly low sales. It was received wisdom in thoughtul Mac circles that Microsoft could kill Apple whenever they wanted simply by cutting Office support, but now they're threatening to I don't feel that worried.

Mac:Office X costs $500 for home users, increasing the cost of a low end system by 50%. In return you get to continue to swap the same old word processing files with friends, which is useful but offers no significant benefit beyond what was offered in Mac:Office 98. An all OS X environment is nice, but it's not worth the $500 MSFT is charging. Moreover, while upgrades are cheaper, they're still pricey and I don't know of any PC to Mac upgrade offers, which makes it harder for Windows users to switch.

OS X has 2.5M users, and Microsoft's sold 300,000 copies of Mac:Office X -- about a 10% penetration rate which at $500 isn't bad. People think Microsoft responded to the threat of the Internet by getting the world to use IE, but the real battle is over HTML, TCP/IP, FTP, SMTP, plaintext, and all the data formats people now spend most of their time with. .doc, .xls, and .ppt are simply less important than they used to be, and "good enough" open office compatibility will reduce Microsoft's power to price these applications high.

So I guess the real reason I'm not worried if Microsoft pulls support for Mac:Office is because I think Mac:Office X (which I get cheap because I'm a student) is the last copy of Microsoft Office I'll ever need to buy.

Saturday, July 13, 2002

Lawyers and Economists Larry Staton Jr takes issue with my "clueless lawyers" post saying
Lawyers do understand economics. Lawyers only want to help musicians if they are their client. Unfortunately for musicians, they do not have enough money to compete with the corporations that enacted CARP, etc... Law, for better or worse, is a business first, advocacy second, idealism last.
I'd make a distinction between understanding where your own self interest lies (which most people manage pretty well) and making policy that benefits society as a whole (where you need to understand economics). And I also lay the blame squarely with an overreaching government--if elected officials offer to sell favors, people are only rational to buy them.

And lastly, I would never critisize a corporate lawyer for not understanding his vocation as just that: a business first, advocacy second, and idealism last. But when their advocacy is *contrary* to their idealism it's clear they're not understanding something, and that something is how economics dictates the world works. Reducing the value of what artists have to sell through "moral rights" copyright ammendments can simply not help artists.

Friday, July 12, 2002

Let's be very clear There's a typically confused Salon piece on Microsoft's Palladium announcement, but MSFTs worked hard to keep the issue confused. It's easy to cut through the Palladium FUD if you think hard about how the technology works. 1) Palladium does not stop viruses. Microsoft is not going to veto code executing on your computer. A virus is code executing on your computer, so this will have no effect. 2) Palladium will not protect privacy. Yes, PGP tokens get exchanged, not form information, but this really doesn't matter -- privacy in the sense people care about the concept (are people reading my email? why do I keep getting spammed?) has nothing to do with Palladium. 3) Palladium is about DRM. Signed content is DRM content, and only that will care whether Palladium is there or not. This will support the content cartels and their Congressional employees.4) Palladium does not help consumers. Consumers don't care whether content is signed or not. Virii and spam will continue unabated. 5) Palladium will not kill open source. Let's be honest here, the only thing that'll kill open source is the US Government, either through some idiot patent policy, or idiot law.

Thursday, July 11, 2002

User led innovation I'm a big fan of Hal Varian's tech economic--it's cold blooded and good at explaining why things work out they way they do. He has a nice Times piece on the negative effects of downstream restrictions management, arguing that since (experimental) user innovation is critical to technological development, limitations on use will retard progress.

Winterspeak back up Blogger ate my template, but everything is back up now. This is one of the few things I dislike about Blogger -- every so often it does something that eats my blog and makes me wish I had a stable, client side application.

Wednesday, July 10, 2002

Clueless lawyers Lawyers, like most people, never studied economics and so don't understand how things happen in the world. But unlike most people, lawyers get to set policy through laws, and this ignorance leads to actual harm. Edward Samuels is just such a lawyer, and he claims that economic justice demands artists get more money for their work and suggests copyright should support that.

Unfortunately, economic justice guarantees that the lot of the artist will always be difficult, as many people want to do it and one unknown artist is interchangable with another. With the laws of supply and demand being what they are, what is offered by many and valued by few will clear at a low price. Any law that gives non-transferable claims to artists will merely lower the price they're paid for their work, as the owners of scarce promotion and distribution resources capture rents just as they always have. If lawyers want to help musicians, they should scrap CARP, dismantle mandatory contract clauses, and investigate the music industry for price fixing and collusion.

Party like it's 1999 We took a tour of IBM's e-Business innovation center in Boston the other day. Its aeron chairs, hipster accessories, and breathless gushes of "creatives" and "technicals" reminded me of Razorfish et al in the bad old days of the Internet boom. I didn't know any of the old-style digital agencies were still around, I thought their senseless tolerance for self-indulgent drivel, lack of business understanding, and plain ole silliness had driven them all to bankrupcy, but it looks like this one managed to find refuge in Big Blue's shaggy hide.

After tolerating a particularly asanine talk on why Chubb Insurance's site should "feel" like Ferrari's, I left. In retrospect, I think part of the reason digital agency's did as well as they did was because they made clients feel "happening" and these pecuniary benefits outweighed any corporate responsibility to customers or shareholders.

Tuesday, July 09, 2002

Why did eBay buy PayPal and Billpoint About six months ago, eBay bought online payment system Billpoint sending rival PayPal's stock price down 15%. Six months later, eBay buys PayPal anyway. Why?

PayPal was clearly the prefered way to pay for goods on eBay, but that alone does not explain why the two should integrate. After all, eBay offers other payment options and would clearly like that complementary function to be commoditized. But the eBay user experience, while not great, could be improved by better integration with PayPal ( combined eBay/PayPal accounts, for example) and owning the company makes that easier to do. eBay initially hoped to do this with the cheaper Billpoint, but when that didn't work out, they stumped up for PayPal.

Another reason is that PayPal's large number of established accounts gave it some pricing power through charging higher service rates, just as eBay's market power for unique items lets it charge higher auction rates, and summing the two markups lead to higher prices than was optimal for the system as a whole. eBay corrected this by buying PayPal and (I'm guessing) will lower the (combined) transaction prices.

The harmful effect created by stacking monopolies on top of each other is called double marginalization and is fixed in a free market by one monopoly buying the other. Essentially, as monopoly markups sum linearly (x), the sum of the harm they cause grows to the square (x^2). By one monopoly buying the other, they can eliminate the double markup and merely charge a (lower, less harmful) single monopoly price that is also more profitable for the monopolist.

Monday, July 08, 2002

License 6.0 I wonder if license 6.0 / XP will be the turning point for Windows. With 6.0, Microsoft kills its last remaining competitor, older versions of itself, and moves towards a world where an organization writes Bill a big check so it can use its computers. Given how user hostile Palladium is, increasing the price through 6.0 encourages people to consider alternative platforms.

Sunday, July 07, 2002

She said yes! I am the happiest man in the world.

Wednesday, July 03, 2002

Will OS X.2 ship early? CNet has a rumor that Jaguar will ship early. This makes sense since at least some customers (such as myself) are waiting for 10.2 to come out before shelling out for a new system, and iMac sales have stalled. Perhaps Apple should not have discussed Jaguar at their World Wide Developer Conference (WWDC)--the stunning performance upgrades and real possibility of having to pay for the upgrade makes it worthwhile to wait until it's released before buying a new system.

Tuesday, July 02, 2002

Palladium I spoke with my excellent friend, who is also the Security expert at Well Regarded IT Analysis Firm, about Microsoft's Palladium plans. The info out there is pretty different from what she told me, but I trust her opinion, so here's the skinny: Palladium works by including a new piece of hardware (called a nub) that stores a public key. The CPU can talk to the nub to see what the public key is and use it to support secure authentication. Note: like Microsoft's CLR, "unsafe" or "unauthenticated" code can be executed, it'll just be "unauthenticated".

This means that locally compiled GPL executables can run after all, as can old "unauthenticated" software. This also means that Palladium will do nothing to protect users from viruses, worms, etc. which will simply autoexecute in Outlook just as they always have. But users don't care about security much anyway.

So what will Palladium actually do? It'll let content producers distribute content that then cannot be redistributed. Palladium will broker a transaction between RIAA servers and your own and transfer encrypted binaries over the network that must have that specific CPU's public key to execute. To be honest, I have no problem with content owners locking up their precious bits ever more tightly, it'll just encourage open alternatives so long as people are allowed to also distribute stuff that's free. And Palladium (as of yet) seems to allow that.

Monday, July 01, 2002

Cringely vs IBM I found an old Cringely article where he critisizes IBM for ignoring, or crushing, innovation. His advice to Palmisano is
IBM is too centralized. Run the divisions like independent companies. Get new financial systems and run those independently, too. Think of IBM as a holding company with lots of business units, each of which has its own profit and ROI goals.
Like almost every business reporter, he has no understanding of business, and like every technology reporter, he has a warped sense of how the technology business works.

The major advantage big tech companies have is their installed base and established distribution channels, not the quality of their products. Innovative "better" technologies fail because they have no distribution or support, not because their tech's any good. By contrast, big companies are good at selling what they already have and selling stuff that's very similar to what they already have, but poor at making good, innovative tech. In fact, it's the very channel specialization that let's them sell so effecitvely that hampers their ability to sell (or notice) anything novel.

The solution to this dilemma is for big companies to stop investing in R&D, which never makes it to the market anyway, and focus on operating efficiency and scouring the market for new tech to sell through existing channels. IBM is world famous for inventing great things and then seeing someone else walk away with all the profit, and I don't think they can ever avoid this fate. Intel, quite conciously, focused on improving CISC production technology and has squeezed more life out of it than the "innovative" RISC set could ever have imagined.

Cringely suggests IBM should become a bunch of small, independent companies under an umbrella organization, apparantly because conglomerates have yet to fall out of favor in PBS land. A bunch of small independent units might as well be small independent companies unless there is some supracontractual benefit to being together, but presumably you'd need cross-group committees to take advantage of these benefits and then suddenly you're back to Big Blue again.