Yes. Apple does have a right to a monopoly on Apple-branded computers. The Coca-Cola corporation has a right to a monopoly on Coke. Nike has a right to a monopoly on Nike running shoes.
The whole purpose of patent, copyright, and trademark is to grant a monopoly on ideas, expressions of ideas, and icons associated with a specific company.
If you want a computer whose OS is just as good as OS X, you have every right in the world to go out and write one. Apple can't do shit to stop you. You can sell it for profit, or just give it away if you want. You might even create a license that requires people who use and modify your code to release their own modifications so that other people can continue to share the wealth.
That's what we call "a competitive market."
Taking work that someone else spent the time and money to create, then using it to compete against them, is called "being a huge flaming asshole"... a concept you've obviously mastered. If you do it in contradiction to the terms of the license -- the one whose validity is defined in terms of the monopoly granted by patent, copyright, or trademark -- that's called "illegal."
There's both a legal and ethical difference between "I'm willing to share everything I have with you," and "I'm willing to share everything you have."
Apple doesn't sell OS X for non-Apple-branded hardware, and Psystar doesn't sell copies of OS X per se. Psystar isn't taking any money out of Apple's pocket in a software market for 'copies of OS X', because no such market exists.
The place where Psystar is taking money out of Apple's pocket is in hardware. The difference between "an Apple-branded computer running OS X" and "a Psystar-branded computer running OS X" is the machine itself, not the OS. Therefore, the damage to Apple is the cost of equivalent hardware less the cost of one retail copy of OS X. Apple's retail computer prices are public knowledge, so adding up the damages would be trivial.
That's only where Psystar's barbed-wire enema begins, though. The big money will be in 'brand dilution'.
Apple's brand is immensely valuable, and Psystar has been trying to redefine it. Apple has always presented itself as a company that sells "the whole widget." They've spent a lot of money in every aspect of their business, from R&D to manufacturing, to sales and support, pushing the idea that a box from Apple is a single unit that Just Works (TM). Then along comes Psystar, telling everyone that Apple is really just a component vendor, whose OS division is just like Microsoft and whose hardware division is just like Dell. If you want an "Apple compatible computer," you can pick and choose pieces from any vendors you want.
A large part of this case revolves around whether Psystar has a legitimate right to tell the whole world what business Apple is in. At present, it looks like the answer to that is a big, fat "No."
If a court rules that Psystar has indeed been blowing smoke out its ass, then the equivalent advertising cost of every square inch of newsprint or second of airtime devoted to covering this case in the public view can be treated as damage to Apple. Every "they can't tell me what to do with the OS after I've bought the box" comment in these threads counts as noise that Apple will have to spend time, effort, and money arguing against. That time, effort, and money are resources Apple could otherwise have spent writing more software, designing new hardware, or otherwise expanding or improving its business. Psystar would have imposed an operating expense on Apple without having any legal right to do so. And following the theory that "if you light the fire, you can be held liable for whatever gets burnt," the court can rule that Psystar owes Apple the cost of cleaning up the mess it created.
Good point, but I personally think Visicalc tipped the scale in Apple's favor.
Visicalc was the killer app that made buying a micro worthwhile, at least in the business market (which was the only 'real' computer market in those days). For a couple thousand bucks, a company manager could buy an Apple ][ and Visicalc and have the power of The Spreadsheet at his disposal. That was a big deal back then, especially since you could run a calculation in real time (read: 5-15 minutes) rather than having to submit a batch job to the computer department and get the results back tomorrow at best.
IBM got interested in the microcomputer market when it learned that middle managers were smuggling Apple ][s into their offices so they could crunch numbers. That presented a challenge to IBM's core market, and by the time anyone admitted that it was happening, it was happening in a fairly big way.
Cloning a PC wouldn't be legal either if IBM hadn't screwed the pooch on getting the first product to market.
First of all, IBM massively underestimated the potential for growth of a microcomputer market. They held the idea that Big Iron was the only true definition of computing, and that things like the Apple ][ were hobbyist toys that would never amount to anything. They made the further mistake of assuming that if a market for 'toy' computers did start to become worthwhile, they'd have plenty of time to develop and ship their own product.
They were wrong in both cases.
When IBM finally decided to sell a PC, they were of the opinion that Apple was 6-12 months from getting a lock on the microcomputer market. If IBM couldn't put a product on the shelves by that time, there wouldn't be much point in trying.
So they tasked an engineer with the job of creating enough of a product to hold a space in the market while the designers put together something really good. Being a good engineer, he did a baseline critical path analysis, and learned that with all the forms, paperwork, and meetings, it would take something like 18 months to ship an empty box with "IBM PC" printed on it. Actually designing a computer to put in the box, shopping for parts suppliers, building an assembly plant, and all those other little details would just push the ship time farther out.
So, faced with the choice between losing a new market entirely or skirting around standard procedures, he proposed a radical plan: design a machine out of off-the-shelf parts, and contract third-party assembly shops to do the construction. That would allow IBM to put a product on the shelves within the 6-12 month deadline, but it would also create an enormous risk: anyone else who wanted to enter the market would be able to do exactly the same thing just as quickly, and just as easily. In fact, it would be even cheaper and easier for the me-too competitors, because they could skip the R&D phase and copy IBM's hardware design more or less verbatim (a process that came to be known as 'cloning').
So they built the whole thing around a chip which could only be sold by IBM: the BIOS.
The BIOS was a computer program burned into ROM. IBM held the copyright on the program, so nobody could legally duplicate that chip. But the BIOS was also tightly integrated with the hardware. Without it, the rest of the computer was just a box of random components. But those components were arranged in such a specific way that it would be hell to try to design a compatible product that wouldn't require IBM's BIOS to run. In one version of the fantasy, IBM wouldn't have to build computers at all, they'd just license BIOS chips to all the other companies that wanted to build hardware.
Then along came a company called Compaq, which reverse-engineered the IBM BIOS, and built a legally clean BIOS of their own from the reverse-engineered spec.
IBM sued, and lost. Compaq's legal team had done its homework on maintaining the 'virginity' of the coders who wrote the cloned BIOS.
At THAT point, IBM lost all control of the PC hardware market. And since their OS had also been outsourced to a little company up in Seattle, they didn't have any hooks left in the product.
So in answer to your question: it's legal to clone a PC because IBM was lumbering, stupid, arrogant, in a big hurry, and not thinking very clearly when it spent tons of money pushing a design into the market that could be ganked away from them almost overnight. In the process, they handed half of their market dominance to Microsoft (whose OS became the only thing that made the hardware a 'PC') and the other half to cloners like Compaq.
Those "consumers who are always right" are the ones who want Hollywood movies in high definition, regardless of what technical encumbrances are baked into the formats and licensing deals. They want music players full of music from the big five's catalogs, and generally don't care about DRM as long as it doesn't get too much in their way.
If consumers were lining up with money in hand to buy DRM-free products, and collectively saying a big "screw you" to the RIAA/MPAA by refusing to buy the latest music and movies as long as they were encumbered by DRM, Apple would no doubt tell the RIAA/MPAA to shove their DRM clauses up their collective ass.
Apple has already held tough on issues like flexible song pricing, and was willing to play chicken with NBC over "flexible video pricing plus bundling," so I think they've earned the right to say they'll act in the consumer's best interest.
If consumers in their millions aren't interested in what you think is 'best', then I guess it just sucks to be you.
--- You are partly right, but surely the issue is, they have no right to restrict what A RETAIL COPY OF OSX is installed on.
Of course they do.
Copyright law says that by default, the creator of a work has all the rights that exist in regard to the created work. By default, no one else has any rights to the work at all.
If anyone but the creator has any rights to the work, those rights have to have been granted by the creator. Period.
If you don't have some kind of document that says the creator has given you a specific right, you don't have that right. Period.
The rights that the creator gives you are called a 'license', and the creator can put any terms into the license he wants, as long as those terms don't break some other law. There's nothing in the law that says those terms have to be fair, or to your liking. 'Fair' is too hard for the law to define, and if you don't like the creator's terms, you have the options of negotiating other terms, or not buying the license and accepting the fact that you have no rights to the work at all.
License terms that tell you what you can do with a product after you've bought it are called 'performance restrictions', and performance restrictions are utterly legal.
The GPL and most OSS licenses are nothing but a big bundle of performance restrictions. You can't take a piece of GPL'd code and use it in a closed-source product by saying, "hey, I bought A RETAIL COPY OF LINUX. You have no right to tell me what I can do with the code after I've bought it."
Apple hates DRM because it takes a shitload of time, money, and effort to design and implement, and an even bigger shitload of time, money, and effort to show 'plausible efforts' to keep it working once it's deployed. Anyone who tries to do DRM simply agrees to climb on a treadmill of trying to stay a fractional step ahead of the people who will break the protection.
Apple hates DRM because DRM is inherently futile. You simply can't build a system that puts both the lock and the key in the hands of the end user, then impose rules on what the user can do with those two pieces.
Apple hates DRM because the longer DRM is allowed to exist, the longer the content cartel will continue to make this massive, futile investment a requirement for any access to content. Apple especially hates DRM because it puts the content owners in a position where they can say, "I don't have to know how it works, or whether it's even possible. I have the power to say what has to be done, and making it happen is your job. And thanks to the laws that we've bought and the contracts we've written, if you don't manage to do the impossible to our satisfaction, we can sue the shit out of you then nail you on criminal charges."
Apple hates DRM because using DRM simply manufactures enemies with the technical knowledge to rip apart any technical measures Apple tries to build. And while defeating DRM may be a socially acceptable goal, a lot of that knowledge can potentially be reapplied to general malware.
Apple hates DRM because it sucks for the user. Remember: Apple doesn't make money licensing its OS to a bunch of OEMs who then try to sell a product to consumers, or with massive, umpty-thousand-seat software licensing deals. Consumer dissatisfaction hits Apple in the pocket much harder than it hits Microsoft. On top of that, Apple sells in the premium-priced segment of the market, where people are willing to say, "if I have to put up with something that sucks, I can buy another product for a whole lot less."
It would take Apple a hell of a lot less effort to make a product that users like a hell of a lot more if they could ditch DRM. Given that less than 1% of all the music on iPods was purchased through the iTunes store, the idea that Apple sees some kind of benefit from consumer lock-in just doesn't scan.
The only upside of DRM is that it gives Apple access to the content cartel's catalog. And tens of millions of consumers voting with their dollars have said that they prefer devices that do have DRM and cartel content over devices free from DRM that don't have cartel content.
So in the long run, it comes down to a question of philosophy versus economics. If you hate DRM so much you won't buy a Mac, iPod, or iPhone, so be it. The few thousand people who agree with you on that score are less valuable to Apple than the 15-20 million who will buy new DRM-encumbered Macs, iPods, and iPhones in the next quarter.
--- The point you make is the exact point all corporations make in order to exploit cheap foreign labor. "Well, their lives sucked, so let us pay them peanuts, then they must be happy"
Right.. it's not like Foxconn and similar companies have dumped billions of dollars into areas that had nothing to offer the modern global economy except human bodies with the capacity to learn skills that would eventually turn them into a labor market.
So the previous standard for accommodation was a shack with a dirt floor and walls made from salvaged materials, and the people are now living in modern, insulated, climate-controlled structures whose building materials alone the local economy never could have produced. Who cares?
So the water came from the nearest creek, which was probably within the leach field of the local outhouses. So the people now get their water from a system which is properly cleaned and treated, and don't even have to carry a 40-pound jar back from the creek to get it. What's a little cholera now and then?
So the buying power of a $450/month salary is a hell of a lot higher in that young and growing economy than it is here in the US. So the people now have access to a range of goods and services beyond the local dreams of avarice twenty years ago. If you listen to the corporate exploiters and their butt-boys, you'd think that's some kind of big deal.
So the local economy now has people with marketable skills, ranging from construction and water treatment to accounting and electronic fabrication. So the workers are earning enough money to leave that local economy and take their skills to another, higher-paying market if they can find one. So the skills will remain even if Foxconn drops out of existence tomorrow. So the local economy will eventually outgrow Foxconn, and Foxconn will eventually have to compete for labor.
It doesn't matter if it's 1992 or 2008, there will always be a thing called 'market strata'.
In very rough terms, the 80-20 rule applies to sales: 20% of the sales generate 80% of the profit. The exact percentages vary, but there's a small number of sales at the top of the market that generate a whole lot of profit, and there's a small number of sales at the bottom of the market that are a dead loss.
That fact colors the distribution of operating system adoption.
Apple sells to the top of the market. Their machines are 'overpriced', meaning competitively priced for features and build quality with the other high end machines out there. But where companies like Dell offer bargain-basement prices for bargain-basement quality machines, Apple doesn't. OS X is therefore tailored to the top end of the market, and while it may not have the infinite tweakability that Linux junkies want, it does have a very high level of fit and polish.
Microsoft sells to the middle of the market. Their high-dollar contracts are "N-thousand seats of an unexciting configuration" deals with major corporations, and they support all those home machines as a way of maintaining their lock on the business market. Windows is as much of an all-things-to-all-people product as Microsoft can reasonably deliver and support in units of tens of millions.
Linux and the various FOSS operating systems sell to the low end of the market. They provide phenomenal support for hardware that's ten years old (i.e.: fully depreciated and thus 'worthless'), and have the DIY capacity to work in a hundred million edge cases that no mass-distribution software house could ever hope to support.
Apple doesn't want to support the "I want all the features and polish of OS X, but don't want to pay for the hardware, and oh, I also want it to have all the skins and options of my personal favorite Linux distro" crowd. By definition, those people aren't going to put any money in Apple's pocket, and supporting them would cost a whole lot more than it was worth. Apple's interest in the middle-of-the-road business market is something they'll probably take in passing where they can get it. The requirements for a business machine haven't changed all that much in the last fifteen years or so, and if they can support the basic feature set with a minor tweak to their default installation, why the hell not?
I don't see Apple ever wanting more than about the top 30% of the market, though. That's where all the money is. I see Apple as far more interested in keeping Microsoft out of the high end market than in trying to expand its own sales into areas where it will have to work harder for less return.
FOSS, on the other hand, is very much interested in moving up into Microsoft's core markets. Those corporate contracts are worth a whole lot to people who normally give code away for free, or to companies that can bolt a paid product on top of all that free-as-in-beer software. And the higher FOSS operating systems can climb into Microsoft's territory, the more developer interest and user feedback they'll attract.
Interestingly, Apple can squeeze Microsoft from both sides by being a good citizen of the FOSS community. Things like WebKit push Microsoft to work harder to protect its core markets from free software encroachment, but also make Apple's technology a much wider foothold than it would get if it were strictly tied to OS X sales.
FOSS is inherently able to support the middle and low ends of the market better than Apple could ever hope to do, and Apple will always be there for people who are willing to spend a few grand for a computer that "just works, right out of the box" if it means they can get on with what they really want to do with the machine.
The way I see it, Microsoft is the one facing the real challenge. It has strong competition from Apple at the high end, and is just as vulnerable to FOSS at the low end as Apple would be, so it needs to find a way to cement its hold on the center of the market.
You have it backwards: Apple makes more money than Dell.
On May 29, 2008, Dell posted quarterly financial results of $16.1 billion in revenue, and $784 million in profit.
On April 23, 2008, Apple posted quarterly financial results of $7.51 billion in revenue, and $1.05 billion in profit.
So.. Dell sold more machines than Apple, and took in a bit more than twice as much cash in sales. But Dell only kept about one dollar for every twenty that came in. Apple, OTOH, kept about one dollar for every seven that came in, and took home about 25% more money than Dell did.
Apple's current market cap (stock price times the number of outstanding shares) is $151.47 billion.
Dell's current market cap is $49.25 billion.
So.. just to drive the message home for the terminally hard of thinking: Apple made about 25% more money than Dell did this quarter (and has done for a while), and Wall Street seems to think Apple is worth about three times as much as Dell.
---- And while the "several days" of base load could be turned into a week or more with the correct engineering, that raises construction costs significantly...
Costs like that are the easiest to justify during planning. It's a one-shot payment with little or no ongoing maintenance cost, and as you said, the things could last damn near forever. Take away the recurring expenses and even 'significantly more expensive' amortizes out to 'almost free' over a very long time.
That 'profit' has to be factored against the $19-per-megawatt tax credit wind power receives, plus other government subsidies.
Don't get me wrong.. I think wind power is worth exploring, and I know that any new technology can use a financial boost to help it ramp up from the lab and test-factory stage to the point where economies of scale kick in and it really starts to make money.
Bottom line, though: wind power hasn't gotten to the point where it's competitive with coal, gas, or oil on its own merits yet.
So financially speaking, there is a valid comparison to be made between wind and cold fusion. Everyone agrees that something is happening in cold fusion reactions, because more power comes out than was pumped in. Researchers just haven't gotten to to the point where enough power comes out to make the reaction self sustaining and energy positive. Everyone agrees that wind power looks like a good idea, but it hasn't gotten to the point where enough money comes out to make it financially self sustaining and profitable. There's a good chance we'll see breakthroughs that put both fields in the black, but we haven't seen them as of today.
Rather than buying resistors and capacitors in simple 1-10-100 values, look into the 'E' series of preferred number values.
Preferred numbers exist to make scaling designs easier and to reduce the error when rounding from 'the value required by the design equations' to 'what I have in inventory'. The roughest official series of preferred numbers for electrical components is called E6: 10, 15, 22, 33, 47, 68. The series jumps up one order of magnitude every six steps, hence the '6' in 'E6'.
Each value in E6 is about 150% of the previous one, which means you can scale a design up or down 50% simply by shifting all the component values one place right or left on the scale. There are similar scales for E12, E24, E48, E96, and E192, each of which has twice as many values as the previous one, and offers twice as fine a grain of scaling potential. Here's a nice reference chart.
The maximum rounding error between any random number from 10 to 100 and the nearest E6 value is about 20%. E12 has a maximum rounding error of about 10%, E24 clamps the error to about 5%, etc.
If you start off buying parts from the unofficial 'E3' scale -- 10, 22, and 47 -- you can get within 10% of any multiple-of-ten value with at most three components: 10, 22, 32(10+22), 42(10+10+22), 47, 57(10+47), 69(22+47), 79(10+22+47), 91(22+22+47). To get closer than 10%, you can add (at most three) components from the next order of magnitude down.
The E series also give you a nice, predefined wishlist for future component purchases. Once you have five or six decades (orders of magnitude) of E3 values in your inventory, you can start filling the gaps with E6 values (15, 33, 68). Then you can move on to the E12 and E24 values if you really want to get serious about it. Note that values from the E24 scale don't carry over to the E48 scale (150 vs 147, 120 vs 121, etc), but I seriously doubt that the issue will ever arise for the average hobbyist.
Apple's approach to back-compatibility has allowed them to migrate across three hardware architectures and a completely different OS architecture without seriously alienating their customer base.
I have a 400 MHz G4 that I bought in 2001. Out of the box it ran OS 9. Now it runs Tiger, and is a perfectly acceptable machine. I don't use it to transcode video, because my 2.4 GHz MBP does that a lot faster, but I don't have any problems surfing the web, reading email, writing code, or playing DVDs. And most of the software I use on the G4 is the same stuff I use on the MBP, installed from the same disks if I couldn't just drag-and-drop the application from one machine to another.
Granted, my G4 can't run Leopard, and I'm sure there's some Intel-only or Leopard-only software out there that it can't run either. But generally speaking, I don't care. The software I actually use on it will probably continue to have 10.4 PPC support for a few more years. When the stuff I use day to day no is no longer supported on that machine, I'll still have a perfectly good fileserver/webserver that I can run headless and control from my laptop via VNC.
I'll probably see a 13-15 year lifespan out of that box running OS and software released by Apple, and then I can turn it into a NetBSD box and expect probably another ten years of life before the hardware is so far out of date that it just isn't worth keeping any more.
Most of our time is wasted getting the first bit from the client to the server.
Most network provider SLAs guarantee a round-trip time of around 50 ms or less between any two backbone routers in the US. Add consumer connection latencies (which tend to suck), non-backbone routers, HTTP connection latencies, process spawning times, and all that, and we can expect to see a delay of maybe 50-75 ms from the moment the first bit leaves the client machine and the moment it can be processed by the server.
Once the first bit actually arrives, we switch from latency calculations to bandwidth calculations. On a 56K modem, it takes about 200 ms to transmit 1K of data. Yes, that's significant if you're living out in the internet-have-not regions, but it's less of a nuisance than the 60-95 images the web designer just had to include in the page. Over 802.11n, 1K of data would go across in about 4-5 ms. In other words, it takes the bandwidth of a 56K modem to make standard connection latencies look insignificant. If we take 802.11g as the limiting bandwidth of an arbitrary connection, we can realistically expect to send 25-30K of data before bandwidth use matches the connection latency.
Translating all this to CPU utilization, 1 ms equals about 1 million instruction cycles for a 1 GHz processor. Pessimistically, you'll burn 100-150 million instruction cycles worth of time just waiting for the connection to open, assuming your AJAX page is itself completely stateless, and maybe another 20-50 million instruction cycles worth of time required to transmit the data.
The actual processing time will be trivial by comparison. That means you can do just about anything to trade CPU time for bandwidth and it will pay off for cases of large data volume. It will pay off for small data volumes as well, if transmission speed is the only thing we care about, but there's really no point shaving another 2 ms off the transmission time when the latency is still 150 ms.
By all means gzip your data before transmitting it, if you feel any performance problems. That will buy you at least 3:1 compression for arbitrary text, and more for text with lots of recurring elements like XML tags. For XML-serialized data structures, gzipping will make most of your XML overhead go away.
Assuming you actually want a reasoned answer, it's because web browsing is an important use-case for most people who use computers these days. The way most people see it, a computer that doesn't have a web browser isn't worth owning.
That means whoever makes the web browser for a given OS has a whole lot of power over that OS.
So.. why should any company with as much money and as many programmers as Microsoft or Apple -- knowing that 'not having a web browser' will kill product sales real fast -- just shrug and say, "oh, we'll let somebody else do that."
Apple and Microsoft develop their own web browsers because they want to make sure the people developing their web browsers have Apple's or Microsoft's interests at heart. No, "nah, we don't feel like adding that feature." No, "yeah the performance is lousy on your OS, but we'd have to reorganize the code all the way to the ground to make it better (aka: Mozilla)." No, "yeah, compared to applications actually made for your OS, the interface looks like we put lipstick and a push-up bra on a bulldog's ass. But we have to make sure our browser looks the same everywhere (also aka: Mozilla)." No, "we don't care if you want to schedule a release date for your new OS, we'll release the next version of our browser when we damn well feel like it." No, "maybe you want to deprecate this part of the API, but we intend to keep using it, so bite us."
It's pretty much the same reason Apple chose to develop Safari on its own rather than sitting around waiting for Microsoft's MBU to get funding for a new version of IE for Mac.
Beyond that, there's the issue of implicit endorsement. A computer without a web browser won't sell, so you have to bundle a browser with the OS to get it onto a consumer's desk. The consumer then makes the (reasonable) assumption that you endorse the software you've bundled with the computer. If the software doesn't work the way the consumer wants, the consumer isn't going to complain to some nebulous open-source project, they're going to complain to YOU. After all, you're the one who just got all that money from them.
Microsoft was clearly talking cow pies when they said that a web browser should be considered an integral part of the OS, but they were right that the web browser is too important to an OS to just farm it off to a third party.
You're confusing 'market share' with 'profitability'.
The 80-20 principle applies to markets. 80% of the profit comes from 20% of the sales, and the remaining 20% of the profit comes from the remaining 80% of sales. Generally speaking, if you sell five machines, you make $80 from one of them, and $20 from the remaining four (which works out to $5 per unit).
That means a single high-end sale is worth 16 times as much as a single low-end sale, or that you can work ten times as hard selling low-end machines and still make less money than someone who concentrates on selling high-end machines.
And that matches the actual, reported revenue figures in the market. Dell is still the big dog when it comes to units shipped. They move five times as many units as Apple. Dell's 2Q profits for 2008 were $733 million, a 46% improvement over last year. Apple's 2Q profit for 2008 was $1.05 billion, up 36% from the previous year.
In other words, Dell is working five times as hard for 3/4 as much money. That's what you're suggesting Apple should do.
Your idea is reasonable in theory, but you have it backwards to reality. In reality, you can assemble a plausible business case to support just about anything you want to do.
You don't tell shareholders, "we won't do business with company X because their CEO spanked our CFO on the golf course, and now anyone who suggests using company X's products risks getting their budget cut," you just talk about the uncertainty of the market with respect to company X's product (which is a gimme, because markets are always uncertain.. just look at the endless stream of iPod Killers hailed by the business press over the past N years), and express doubt that the returns from such a project would be sufficient to justify the risk. Then you puff up some other project that costs enough that you can't afford to do that and make a deal with company X.
I've seen multimillion-dollar deals decided on which director could outstare the other at the conference table. I have a friend who's a business consultant whose main line of business involves giving companies an excuse to abandon the project that everyone knows to be a waste of money, but which no one can shit-can directly for political reasons.
Business is a lot more visceral than most people would like to think. Senior executives end up where they are for having reliably good taste (Linus Torvalds once described himself as "CVS with taste") more than good business-analytical skills. That environment leaves plenty of room for pissing matches, vendettas, and feuds.
It probably doesn't hurt that Leopard is a fully certified Unix.
Linux has a lot of good properties, but full Unix certification has never been a major part of the plan, for philosophical, practical, and financial reasons. But some government contracts require Posix or full-Unix certification, so occasionally the ability to tick off those checkboxes matters.
I've always found it amusing when people use the 'multiple-button mouse' thing as an indication of superior design.
Sure, extra buttons can be convenient. You can use them as shortcuts to operations that would take more effort another way. I don't buy the idea that right-clicking to open a file is more 'efficient' than double-clicking (show me some numbers for the actual time-savings over a year or two), but if that model of interaction happens to float your boat, fine.
The problem occurs when people start building things that require multiple buttons. That creates a hardware dependency that isn't intrinsic to the problem space of the software itself, and boils down to programmers enforcing their notion of 'necessary hardware' through code.
It takes hard work and discipline to build a program that works well with only a single mouse button. You can't get away with putting a zillion features in the right-click contextual menu then burying their single-click equivalents deep in sub-menus or sub-panes of preference panels in applications twelve clicks away.. if you offer single-click equivalents at all.
Apple designs for good usability with a single-button mouse, then allows you to trick it out with extra buttons if you want to. The baseline is "good" and you can improve it from there. Interfaces that require multiple-button mice can be good if you have the hardware their designers want you to have, but their performance often degrades to "sucks rocks" or "nonfunctional" if you don't.
Of course, people adapt to whatever environment they're in. If you get used to the environment where right/middle-click menus are easy and everything else is harder if not impossible, you learn to rate software in terms of what you can do through the mouse. Fewer buttons equals less-capable software, and moving functionality away from the mouse equals frustration.
That's one of the "Aha!" moments I've seen from Windows users who move to the Mac. They don't recognize it as a change in I/O models, because they're still using menus and mouse clicks, but they suddenly realize that things outside a contextual menu aren't necessarily hard to reach. Once they get comfortable with the new I/O model, they find themselves pleasantly surprised by it. What they expected to be frustrating ends up being fairly simple.
---- Small time coders don't need protection in a world without patents (so long as you also prevent cartels and monopolies), because it's very easy to break into a market
And once the small guys break into a market, or more correctly, create a market by building a customer base and finding a sustainable way to generate revenue profitably, the big guys will copy the idea wholesale, throw $50 million in hardware and staff at the problem, and drive the little guys out of the market they created.
That wording is purely defensive for Apple. It more or less says you can't lock Apple out of a given application market just by dropping a quick beta into the iPhone Apps store.
That last bit about "reasonable patents and copyrights" says you still own your code, and Apple can't use it directly without licensing it. Sure, they can spend some of their own development resources writing their own version of a program if yours happens to become popular, but so can every other software house out there.
Point 1: Google's been collecting photos since 2006 and still hasn't completed a first pass.
Point 2: The odds of anyone seeing any specific image already on Google Streets are vanishingly small.
Point 3: Even if people appear in the photos, they aren't identified by name.
Point 4: The vast majority of places where a person can be photographed are neither incriminating nor embarrassing.
That's a far cry from Orwellian surveillance. As for deterrent effect, I could spend a week standing on the corner with a rocket launcher in one hand and a big sign that says, "I'm going to blow up a bank vault," in the other, and still have a better chance of winning the lottery than getting caught because someone caught me through Google Street.
The N-nines model is a fast and easy way to compare order-of-magnitude differences between existing networks, but it says almost nothing meaningful about actual usage or the perception of uptime from a user's perspective.
Let's look at the numbers: 99.9% uptime translates to about 9 hours of unscheduled downtime a year. That can be one 9-hour block once a year, 36 minutes per day, 1.5 minutes per hour, 1.5 seconds per minute, or one dropped packet per thousand. Sure, it's easy to spot a 9-hour blackout, but as the slices of downtime get thinner, they get harder to notice at all, or to identify as USD specifically.
99.999% uptime translates to about 5 minutes of USD per year, and is of questionable value. You can't identify a network outage, call in a complaint, and get the issue resolved in the given timeframe. 99.9999%? It is to laugh. You can't even look up the tech support phone number without blowing your downtime budget for the year. Get hit by a rolling blackout for an hour? Kiss your downtime budget goodbye for the next 120 years.
Getting back to 99.9% uptime, let's move on to standard utilization patterns. USD really only becomes an issue if people notice it.. nobody cares if an incoming piece of email got delayed by 30 seconds at the MTU, but they do get testy if they can't load their webpages. But web surfing only uses 1-2 seconds of bandwidth per minute anyway.
If we have 2 seconds of usage and 2 seconds of downtime per minute, the odds of a collision are around 15:1 with an average overlap of 1 second when a collision does happen. Simply interleaving usage and downtime that way increases the perceived uptime by an order of magnitude since 90% of the outages happen when no one is actually using the network. And larger blocks of downtime get lost in larger blocks of non-utilization exactly the same way.. who cares about a half hour of downtime from 0300 to 0330 when no one in your company is actually in the building and using the network?
Granted, if you have higher utilization you'll have a better chance of hitting a chunk of downtime, but you'll also have higher chances of queuing latency within your own use patterns. If you're already using 99% of your bandwidth, you can't just plunk in one more job and expect it to run immediately. It has to wait for that 1% of space no one else is currently using. And when you get to that point, it's really time to consider buying a bigger pipe anyway.
And that brings us to the main point: People don't buy network connectivity in absolute terms. They buy capacity, and the capacity they buy is scaled to what they think of as acceptable peak usage. "Acceptable peak usage" is a subjective thing, and nobody makes subjective judgements with 99.999% precision.
Slashdot Mirror
Feeding the troll, I know, but what the hell..
Yes. Apple does have a right to a monopoly on Apple-branded computers. The Coca-Cola corporation has a right to a monopoly on Coke. Nike has a right to a monopoly on Nike running shoes.
The whole purpose of patent, copyright, and trademark is to grant a monopoly on ideas, expressions of ideas, and icons associated with a specific company.
If you want a computer whose OS is just as good as OS X, you have every right in the world to go out and write one. Apple can't do shit to stop you. You can sell it for profit, or just give it away if you want. You might even create a license that requires people who use and modify your code to release their own modifications so that other people can continue to share the wealth.
That's what we call "a competitive market."
Taking work that someone else spent the time and money to create, then using it to compete against them, is called "being a huge flaming asshole" ... a concept you've obviously mastered. If you do it in contradiction to the terms of the license -- the one whose validity is defined in terms of the monopoly granted by patent, copyright, or trademark -- that's called "illegal."
There's both a legal and ethical difference between "I'm willing to share everything I have with you," and "I'm willing to share everything you have."
The unit cost of OS X is irrelevant.
Apple doesn't sell OS X for non-Apple-branded hardware, and Psystar doesn't sell copies of OS X per se. Psystar isn't taking any money out of Apple's pocket in a software market for 'copies of OS X', because no such market exists.
The place where Psystar is taking money out of Apple's pocket is in hardware. The difference between "an Apple-branded computer running OS X" and "a Psystar-branded computer running OS X" is the machine itself, not the OS. Therefore, the damage to Apple is the cost of equivalent hardware less the cost of one retail copy of OS X. Apple's retail computer prices are public knowledge, so adding up the damages would be trivial.
That's only where Psystar's barbed-wire enema begins, though. The big money will be in 'brand dilution'.
Apple's brand is immensely valuable, and Psystar has been trying to redefine it. Apple has always presented itself as a company that sells "the whole widget." They've spent a lot of money in every aspect of their business, from R&D to manufacturing, to sales and support, pushing the idea that a box from Apple is a single unit that Just Works (TM). Then along comes Psystar, telling everyone that Apple is really just a component vendor, whose OS division is just like Microsoft and whose hardware division is just like Dell. If you want an "Apple compatible computer," you can pick and choose pieces from any vendors you want.
A large part of this case revolves around whether Psystar has a legitimate right to tell the whole world what business Apple is in. At present, it looks like the answer to that is a big, fat "No."
If a court rules that Psystar has indeed been blowing smoke out its ass, then the equivalent advertising cost of every square inch of newsprint or second of airtime devoted to covering this case in the public view can be treated as damage to Apple. Every "they can't tell me what to do with the OS after I've bought the box" comment in these threads counts as noise that Apple will have to spend time, effort, and money arguing against. That time, effort, and money are resources Apple could otherwise have spent writing more software, designing new hardware, or otherwise expanding or improving its business. Psystar would have imposed an operating expense on Apple without having any legal right to do so. And following the theory that "if you light the fire, you can be held liable for whatever gets burnt," the court can rule that Psystar owes Apple the cost of cleaning up the mess it created.
Good point, but I personally think Visicalc tipped the scale in Apple's favor.
Visicalc was the killer app that made buying a micro worthwhile, at least in the business market (which was the only 'real' computer market in those days). For a couple thousand bucks, a company manager could buy an Apple ][ and Visicalc and have the power of The Spreadsheet at his disposal. That was a big deal back then, especially since you could run a calculation in real time (read: 5-15 minutes) rather than having to submit a batch job to the computer department and get the results back tomorrow at best.
IBM got interested in the microcomputer market when it learned that middle managers were smuggling Apple ][s into their offices so they could crunch numbers. That presented a challenge to IBM's core market, and by the time anyone admitted that it was happening, it was happening in a fairly big way.
Cloning a PC wouldn't be legal either if IBM hadn't screwed the pooch on getting the first product to market.
First of all, IBM massively underestimated the potential for growth of a microcomputer market. They held the idea that Big Iron was the only true definition of computing, and that things like the Apple ][ were hobbyist toys that would never amount to anything. They made the further mistake of assuming that if a market for 'toy' computers did start to become worthwhile, they'd have plenty of time to develop and ship their own product.
They were wrong in both cases.
When IBM finally decided to sell a PC, they were of the opinion that Apple was 6-12 months from getting a lock on the microcomputer market. If IBM couldn't put a product on the shelves by that time, there wouldn't be much point in trying.
So they tasked an engineer with the job of creating enough of a product to hold a space in the market while the designers put together something really good. Being a good engineer, he did a baseline critical path analysis, and learned that with all the forms, paperwork, and meetings, it would take something like 18 months to ship an empty box with "IBM PC" printed on it. Actually designing a computer to put in the box, shopping for parts suppliers, building an assembly plant, and all those other little details would just push the ship time farther out.
So, faced with the choice between losing a new market entirely or skirting around standard procedures, he proposed a radical plan: design a machine out of off-the-shelf parts, and contract third-party assembly shops to do the construction. That would allow IBM to put a product on the shelves within the 6-12 month deadline, but it would also create an enormous risk: anyone else who wanted to enter the market would be able to do exactly the same thing just as quickly, and just as easily. In fact, it would be even cheaper and easier for the me-too competitors, because they could skip the R&D phase and copy IBM's hardware design more or less verbatim (a process that came to be known as 'cloning').
So they built the whole thing around a chip which could only be sold by IBM: the BIOS.
The BIOS was a computer program burned into ROM. IBM held the copyright on the program, so nobody could legally duplicate that chip. But the BIOS was also tightly integrated with the hardware. Without it, the rest of the computer was just a box of random components. But those components were arranged in such a specific way that it would be hell to try to design a compatible product that wouldn't require IBM's BIOS to run. In one version of the fantasy, IBM wouldn't have to build computers at all, they'd just license BIOS chips to all the other companies that wanted to build hardware.
Then along came a company called Compaq, which reverse-engineered the IBM BIOS, and built a legally clean BIOS of their own from the reverse-engineered spec.
IBM sued, and lost. Compaq's legal team had done its homework on maintaining the 'virginity' of the coders who wrote the cloned BIOS.
At THAT point, IBM lost all control of the PC hardware market. And since their OS had also been outsourced to a little company up in Seattle, they didn't have any hooks left in the product.
So in answer to your question: it's legal to clone a PC because IBM was lumbering, stupid, arrogant, in a big hurry, and not thinking very clearly when it spent tons of money pushing a design into the market that could be ganked away from them almost overnight. In the process, they handed half of their market dominance to Microsoft (whose OS became the only thing that made the hardware a 'PC') and the other half to cloners like Compaq.
Those "consumers who are always right" are the ones who want Hollywood movies in high definition, regardless of what technical encumbrances are baked into the formats and licensing deals. They want music players full of music from the big five's catalogs, and generally don't care about DRM as long as it doesn't get too much in their way.
If consumers were lining up with money in hand to buy DRM-free products, and collectively saying a big "screw you" to the RIAA/MPAA by refusing to buy the latest music and movies as long as they were encumbered by DRM, Apple would no doubt tell the RIAA/MPAA to shove their DRM clauses up their collective ass.
Apple has already held tough on issues like flexible song pricing, and was willing to play chicken with NBC over "flexible video pricing plus bundling," so I think they've earned the right to say they'll act in the consumer's best interest.
If consumers in their millions aren't interested in what you think is 'best', then I guess it just sucks to be you.
--- You are partly right, but surely the issue is, they have no right to restrict what A RETAIL COPY OF OSX is installed on.
Of course they do.
Copyright law says that by default, the creator of a work has all the rights that exist in regard to the created work. By default, no one else has any rights to the work at all.
If anyone but the creator has any rights to the work, those rights have to have been granted by the creator. Period.
If you don't have some kind of document that says the creator has given you a specific right, you don't have that right. Period.
The rights that the creator gives you are called a 'license', and the creator can put any terms into the license he wants, as long as those terms don't break some other law. There's nothing in the law that says those terms have to be fair, or to your liking. 'Fair' is too hard for the law to define, and if you don't like the creator's terms, you have the options of negotiating other terms, or not buying the license and accepting the fact that you have no rights to the work at all.
License terms that tell you what you can do with a product after you've bought it are called 'performance restrictions', and performance restrictions are utterly legal.
The GPL and most OSS licenses are nothing but a big bundle of performance restrictions. You can't take a piece of GPL'd code and use it in a closed-source product by saying, "hey, I bought A RETAIL COPY OF LINUX. You have no right to tell me what I can do with the code after I've bought it."
You're joking, right?
Apple hates DRM because it takes a shitload of time, money, and effort to design and implement, and an even bigger shitload of time, money, and effort to show 'plausible efforts' to keep it working once it's deployed. Anyone who tries to do DRM simply agrees to climb on a treadmill of trying to stay a fractional step ahead of the people who will break the protection.
Apple hates DRM because DRM is inherently futile. You simply can't build a system that puts both the lock and the key in the hands of the end user, then impose rules on what the user can do with those two pieces.
Apple hates DRM because the longer DRM is allowed to exist, the longer the content cartel will continue to make this massive, futile investment a requirement for any access to content. Apple especially hates DRM because it puts the content owners in a position where they can say, "I don't have to know how it works, or whether it's even possible. I have the power to say what has to be done, and making it happen is your job. And thanks to the laws that we've bought and the contracts we've written, if you don't manage to do the impossible to our satisfaction, we can sue the shit out of you then nail you on criminal charges."
Apple hates DRM because using DRM simply manufactures enemies with the technical knowledge to rip apart any technical measures Apple tries to build. And while defeating DRM may be a socially acceptable goal, a lot of that knowledge can potentially be reapplied to general malware.
Apple hates DRM because it sucks for the user. Remember: Apple doesn't make money licensing its OS to a bunch of OEMs who then try to sell a product to consumers, or with massive, umpty-thousand-seat software licensing deals. Consumer dissatisfaction hits Apple in the pocket much harder than it hits Microsoft. On top of that, Apple sells in the premium-priced segment of the market, where people are willing to say, "if I have to put up with something that sucks, I can buy another product for a whole lot less."
It would take Apple a hell of a lot less effort to make a product that users like a hell of a lot more if they could ditch DRM. Given that less than 1% of all the music on iPods was purchased through the iTunes store, the idea that Apple sees some kind of benefit from consumer lock-in just doesn't scan.
The only upside of DRM is that it gives Apple access to the content cartel's catalog. And tens of millions of consumers voting with their dollars have said that they prefer devices that do have DRM and cartel content over devices free from DRM that don't have cartel content.
So in the long run, it comes down to a question of philosophy versus economics. If you hate DRM so much you won't buy a Mac, iPod, or iPhone, so be it. The few thousand people who agree with you on that score are less valuable to Apple than the 15-20 million who will buy new DRM-encumbered Macs, iPods, and iPhones in the next quarter.
--- The point you make is the exact point all corporations make in order to exploit cheap foreign labor. "Well, their lives sucked, so let us pay them peanuts, then they must be happy"
Right.. it's not like Foxconn and similar companies have dumped billions of dollars into areas that had nothing to offer the modern global economy except human bodies with the capacity to learn skills that would eventually turn them into a labor market.
So the previous standard for accommodation was a shack with a dirt floor and walls made from salvaged materials, and the people are now living in modern, insulated, climate-controlled structures whose building materials alone the local economy never could have produced. Who cares?
So the water came from the nearest creek, which was probably within the leach field of the local outhouses. So the people now get their water from a system which is properly cleaned and treated, and don't even have to carry a 40-pound jar back from the creek to get it. What's a little cholera now and then?
So the buying power of a $450/month salary is a hell of a lot higher in that young and growing economy than it is here in the US. So the people now have access to a range of goods and services beyond the local dreams of avarice twenty years ago. If you listen to the corporate exploiters and their butt-boys, you'd think that's some kind of big deal.
So the local economy now has people with marketable skills, ranging from construction and water treatment to accounting and electronic fabrication. So the workers are earning enough money to leave that local economy and take their skills to another, higher-paying market if they can find one. So the skills will remain even if Foxconn drops out of existence tomorrow. So the local economy will eventually outgrow Foxconn, and Foxconn will eventually have to compete for labor.
These are slaves. This is exploitation.
It doesn't matter if it's 1992 or 2008, there will always be a thing called 'market strata'.
In very rough terms, the 80-20 rule applies to sales: 20% of the sales generate 80% of the profit. The exact percentages vary, but there's a small number of sales at the top of the market that generate a whole lot of profit, and there's a small number of sales at the bottom of the market that are a dead loss.
That fact colors the distribution of operating system adoption.
Apple sells to the top of the market. Their machines are 'overpriced', meaning competitively priced for features and build quality with the other high end machines out there. But where companies like Dell offer bargain-basement prices for bargain-basement quality machines, Apple doesn't. OS X is therefore tailored to the top end of the market, and while it may not have the infinite tweakability that Linux junkies want, it does have a very high level of fit and polish.
Microsoft sells to the middle of the market. Their high-dollar contracts are "N-thousand seats of an unexciting configuration" deals with major corporations, and they support all those home machines as a way of maintaining their lock on the business market. Windows is as much of an all-things-to-all-people product as Microsoft can reasonably deliver and support in units of tens of millions.
Linux and the various FOSS operating systems sell to the low end of the market. They provide phenomenal support for hardware that's ten years old (i.e.: fully depreciated and thus 'worthless'), and have the DIY capacity to work in a hundred million edge cases that no mass-distribution software house could ever hope to support.
Apple doesn't want to support the "I want all the features and polish of OS X, but don't want to pay for the hardware, and oh, I also want it to have all the skins and options of my personal favorite Linux distro" crowd. By definition, those people aren't going to put any money in Apple's pocket, and supporting them would cost a whole lot more than it was worth. Apple's interest in the middle-of-the-road business market is something they'll probably take in passing where they can get it. The requirements for a business machine haven't changed all that much in the last fifteen years or so, and if they can support the basic feature set with a minor tweak to their default installation, why the hell not?
I don't see Apple ever wanting more than about the top 30% of the market, though. That's where all the money is. I see Apple as far more interested in keeping Microsoft out of the high end market than in trying to expand its own sales into areas where it will have to work harder for less return.
FOSS, on the other hand, is very much interested in moving up into Microsoft's core markets. Those corporate contracts are worth a whole lot to people who normally give code away for free, or to companies that can bolt a paid product on top of all that free-as-in-beer software. And the higher FOSS operating systems can climb into Microsoft's territory, the more developer interest and user feedback they'll attract.
Interestingly, Apple can squeeze Microsoft from both sides by being a good citizen of the FOSS community. Things like WebKit push Microsoft to work harder to protect its core markets from free software encroachment, but also make Apple's technology a much wider foothold than it would get if it were strictly tied to OS X sales.
FOSS is inherently able to support the middle and low ends of the market better than Apple could ever hope to do, and Apple will always be there for people who are willing to spend a few grand for a computer that "just works, right out of the box" if it means they can get on with what they really want to do with the machine.
The way I see it, Microsoft is the one facing the real challenge. It has strong competition from Apple at the high end, and is just as vulnerable to FOSS at the low end as Apple would be, so it needs to find a way to cement its hold on the center of the market.
You have it backwards: Apple makes more money than Dell.
On May 29, 2008, Dell posted quarterly financial results of $16.1 billion in revenue, and $784 million in profit.
On April 23, 2008, Apple posted quarterly financial results of $7.51 billion in revenue, and $1.05 billion in profit.
So.. Dell sold more machines than Apple, and took in a bit more than twice as much cash in sales. But Dell only kept about one dollar for every twenty that came in. Apple, OTOH, kept about one dollar for every seven that came in, and took home about 25% more money than Dell did.
Apple's current market cap (stock price times the number of outstanding shares) is $151.47 billion.
Dell's current market cap is $49.25 billion.
So.. just to drive the message home for the terminally hard of thinking: Apple made about 25% more money than Dell did this quarter (and has done for a while), and Wall Street seems to think Apple is worth about three times as much as Dell.
---- And while the "several days" of base load could be turned into a week or more with the correct engineering, that raises construction costs significantly...
Costs like that are the easiest to justify during planning. It's a one-shot payment with little or no ongoing maintenance cost, and as you said, the things could last damn near forever. Take away the recurring expenses and even 'significantly more expensive' amortizes out to 'almost free' over a very long time.
That 'profit' has to be factored against the $19-per-megawatt tax credit wind power receives, plus other government subsidies.
Don't get me wrong.. I think wind power is worth exploring, and I know that any new technology can use a financial boost to help it ramp up from the lab and test-factory stage to the point where economies of scale kick in and it really starts to make money.
Bottom line, though: wind power hasn't gotten to the point where it's competitive with coal, gas, or oil on its own merits yet.
So financially speaking, there is a valid comparison to be made between wind and cold fusion. Everyone agrees that something is happening in cold fusion reactions, because more power comes out than was pumped in. Researchers just haven't gotten to to the point where enough power comes out to make the reaction self sustaining and energy positive. Everyone agrees that wind power looks like a good idea, but it hasn't gotten to the point where enough money comes out to make it financially self sustaining and profitable. There's a good chance we'll see breakthroughs that put both fields in the black, but we haven't seen them as of today.
Rather than buying resistors and capacitors in simple 1-10-100 values, look into the 'E' series of preferred number values.
Preferred numbers exist to make scaling designs easier and to reduce the error when rounding from 'the value required by the design equations' to 'what I have in inventory'. The roughest official series of preferred numbers for electrical components is called E6: 10, 15, 22, 33, 47, 68. The series jumps up one order of magnitude every six steps, hence the '6' in 'E6'.
Each value in E6 is about 150% of the previous one, which means you can scale a design up or down 50% simply by shifting all the component values one place right or left on the scale. There are similar scales for E12, E24, E48, E96, and E192, each of which has twice as many values as the previous one, and offers twice as fine a grain of scaling potential. Here's a nice reference chart.
The maximum rounding error between any random number from 10 to 100 and the nearest E6 value is about 20%. E12 has a maximum rounding error of about 10%, E24 clamps the error to about 5%, etc.
If you start off buying parts from the unofficial 'E3' scale -- 10, 22, and 47 -- you can get within 10% of any multiple-of-ten value with at most three components: 10, 22, 32(10+22), 42(10+10+22), 47, 57(10+47), 69(22+47), 79(10+22+47), 91(22+22+47). To get closer than 10%, you can add (at most three) components from the next order of magnitude down.
The E series also give you a nice, predefined wishlist for future component purchases. Once you have five or six decades (orders of magnitude) of E3 values in your inventory, you can start filling the gaps with E6 values (15, 33, 68). Then you can move on to the E12 and E24 values if you really want to get serious about it. Note that values from the E24 scale don't carry over to the E48 scale (150 vs 147, 120 vs 121, etc), but I seriously doubt that the issue will ever arise for the average hobbyist.
Apple's approach to back-compatibility has allowed them to migrate across three hardware architectures and a completely different OS architecture without seriously alienating their customer base.
I have a 400 MHz G4 that I bought in 2001. Out of the box it ran OS 9. Now it runs Tiger, and is a perfectly acceptable machine. I don't use it to transcode video, because my 2.4 GHz MBP does that a lot faster, but I don't have any problems surfing the web, reading email, writing code, or playing DVDs. And most of the software I use on the G4 is the same stuff I use on the MBP, installed from the same disks if I couldn't just drag-and-drop the application from one machine to another.
Granted, my G4 can't run Leopard, and I'm sure there's some Intel-only or Leopard-only software out there that it can't run either. But generally speaking, I don't care. The software I actually use on it will probably continue to have 10.4 PPC support for a few more years. When the stuff I use day to day no is no longer supported on that machine, I'll still have a perfectly good fileserver/webserver that I can run headless and control from my laptop via VNC.
I'll probably see a 13-15 year lifespan out of that box running OS and software released by Apple, and then I can turn it into a NetBSD box and expect probably another ten years of life before the hardware is so far out of date that it just isn't worth keeping any more.
That ain't too bad.
Most of our time is wasted getting the first bit from the client to the server.
Most network provider SLAs guarantee a round-trip time of around 50 ms or less between any two backbone routers in the US. Add consumer connection latencies (which tend to suck), non-backbone routers, HTTP connection latencies, process spawning times, and all that, and we can expect to see a delay of maybe 50-75 ms from the moment the first bit leaves the client machine and the moment it can be processed by the server.
Once the first bit actually arrives, we switch from latency calculations to bandwidth calculations. On a 56K modem, it takes about 200 ms to transmit 1K of data. Yes, that's significant if you're living out in the internet-have-not regions, but it's less of a nuisance than the 60-95 images the web designer just had to include in the page. Over 802.11n, 1K of data would go across in about 4-5 ms. In other words, it takes the bandwidth of a 56K modem to make standard connection latencies look insignificant. If we take 802.11g as the limiting bandwidth of an arbitrary connection, we can realistically expect to send 25-30K of data before bandwidth use matches the connection latency.
Translating all this to CPU utilization, 1 ms equals about 1 million instruction cycles for a 1 GHz processor. Pessimistically, you'll burn 100-150 million instruction cycles worth of time just waiting for the connection to open, assuming your AJAX page is itself completely stateless, and maybe another 20-50 million instruction cycles worth of time required to transmit the data.
The actual processing time will be trivial by comparison. That means you can do just about anything to trade CPU time for bandwidth and it will pay off for cases of large data volume. It will pay off for small data volumes as well, if transmission speed is the only thing we care about, but there's really no point shaving another 2 ms off the transmission time when the latency is still 150 ms.
By all means gzip your data before transmitting it, if you feel any performance problems. That will buy you at least 3:1 compression for arbitrary text, and more for text with lots of recurring elements like XML tags. For XML-serialized data structures, gzipping will make most of your XML overhead go away.
Assuming you actually want a reasoned answer, it's because web browsing is an important use-case for most people who use computers these days. The way most people see it, a computer that doesn't have a web browser isn't worth owning.
That means whoever makes the web browser for a given OS has a whole lot of power over that OS.
So.. why should any company with as much money and as many programmers as Microsoft or Apple -- knowing that 'not having a web browser' will kill product sales real fast -- just shrug and say, "oh, we'll let somebody else do that."
Apple and Microsoft develop their own web browsers because they want to make sure the people developing their web browsers have Apple's or Microsoft's interests at heart. No, "nah, we don't feel like adding that feature." No, "yeah the performance is lousy on your OS, but we'd have to reorganize the code all the way to the ground to make it better (aka: Mozilla)." No, "yeah, compared to applications actually made for your OS, the interface looks like we put lipstick and a push-up bra on a bulldog's ass. But we have to make sure our browser looks the same everywhere (also aka: Mozilla)." No, "we don't care if you want to schedule a release date for your new OS, we'll release the next version of our browser when we damn well feel like it." No, "maybe you want to deprecate this part of the API, but we intend to keep using it, so bite us."
It's pretty much the same reason Apple chose to develop Safari on its own rather than sitting around waiting for Microsoft's MBU to get funding for a new version of IE for Mac.
Beyond that, there's the issue of implicit endorsement. A computer without a web browser won't sell, so you have to bundle a browser with the OS to get it onto a consumer's desk. The consumer then makes the (reasonable) assumption that you endorse the software you've bundled with the computer. If the software doesn't work the way the consumer wants, the consumer isn't going to complain to some nebulous open-source project, they're going to complain to YOU. After all, you're the one who just got all that money from them.
Microsoft was clearly talking cow pies when they said that a web browser should be considered an integral part of the OS, but they were right that the web browser is too important to an OS to just farm it off to a third party.
You're confusing 'market share' with 'profitability'.
The 80-20 principle applies to markets. 80% of the profit comes from 20% of the sales, and the remaining 20% of the profit comes from the remaining 80% of sales. Generally speaking, if you sell five machines, you make $80 from one of them, and $20 from the remaining four (which works out to $5 per unit).
That means a single high-end sale is worth 16 times as much as a single low-end sale, or that you can work ten times as hard selling low-end machines and still make less money than someone who concentrates on selling high-end machines.
And that matches the actual, reported revenue figures in the market. Dell is still the big dog when it comes to units shipped. They move five times as many units as Apple. Dell's 2Q profits for 2008 were $733 million, a 46% improvement over last year. Apple's 2Q profit for 2008 was $1.05 billion, up 36% from the previous year.
In other words, Dell is working five times as hard for 3/4 as much money. That's what you're suggesting Apple should do.
---- If Apple gets big into the corporate market, their ads are going to have to change.
Been there. Done that. Back in 2006:
http://www.apple.com/getamac/ads/
Look for the one titled "Self Pity".
Your idea is reasonable in theory, but you have it backwards to reality. In reality, you can assemble a plausible business case to support just about anything you want to do.
You don't tell shareholders, "we won't do business with company X because their CEO spanked our CFO on the golf course, and now anyone who suggests using company X's products risks getting their budget cut," you just talk about the uncertainty of the market with respect to company X's product (which is a gimme, because markets are always uncertain.. just look at the endless stream of iPod Killers hailed by the business press over the past N years), and express doubt that the returns from such a project would be sufficient to justify the risk. Then you puff up some other project that costs enough that you can't afford to do that and make a deal with company X.
I've seen multimillion-dollar deals decided on which director could outstare the other at the conference table. I have a friend who's a business consultant whose main line of business involves giving companies an excuse to abandon the project that everyone knows to be a waste of money, but which no one can shit-can directly for political reasons.
Business is a lot more visceral than most people would like to think. Senior executives end up where they are for having reliably good taste (Linus Torvalds once described himself as "CVS with taste") more than good business-analytical skills. That environment leaves plenty of room for pissing matches, vendettas, and feuds.
It probably doesn't hurt that Leopard is a fully certified Unix.
Linux has a lot of good properties, but full Unix certification has never been a major part of the plan, for philosophical, practical, and financial reasons. But some government contracts require Posix or full-Unix certification, so occasionally the ability to tick off those checkboxes matters.
I've always found it amusing when people use the 'multiple-button mouse' thing as an indication of superior design.
Sure, extra buttons can be convenient. You can use them as shortcuts to operations that would take more effort another way. I don't buy the idea that right-clicking to open a file is more 'efficient' than double-clicking (show me some numbers for the actual time-savings over a year or two), but if that model of interaction happens to float your boat, fine.
The problem occurs when people start building things that require multiple buttons. That creates a hardware dependency that isn't intrinsic to the problem space of the software itself, and boils down to programmers enforcing their notion of 'necessary hardware' through code.
It takes hard work and discipline to build a program that works well with only a single mouse button. You can't get away with putting a zillion features in the right-click contextual menu then burying their single-click equivalents deep in sub-menus or sub-panes of preference panels in applications twelve clicks away.. if you offer single-click equivalents at all.
Apple designs for good usability with a single-button mouse, then allows you to trick it out with extra buttons if you want to. The baseline is "good" and you can improve it from there. Interfaces that require multiple-button mice can be good if you have the hardware their designers want you to have, but their performance often degrades to "sucks rocks" or "nonfunctional" if you don't.
Of course, people adapt to whatever environment they're in. If you get used to the environment where right/middle-click menus are easy and everything else is harder if not impossible, you learn to rate software in terms of what you can do through the mouse. Fewer buttons equals less-capable software, and moving functionality away from the mouse equals frustration.
That's one of the "Aha!" moments I've seen from Windows users who move to the Mac. They don't recognize it as a change in I/O models, because they're still using menus and mouse clicks, but they suddenly realize that things outside a contextual menu aren't necessarily hard to reach. Once they get comfortable with the new I/O model, they find themselves pleasantly surprised by it. What they expected to be frustrating ends up being fairly simple.
---- Small time coders don't need protection in a world without patents (so long as you also prevent cartels and monopolies), because it's very easy to break into a market
...
And once the small guys break into a market, or more correctly, create a market by building a customer base and finding a sustainable way to generate revenue profitably, the big guys will copy the idea wholesale, throw $50 million in hardware and staff at the problem, and drive the little guys out of the market they created.
There was once this company called 'Netscape'
That wording is purely defensive for Apple. It more or less says you can't lock Apple out of a given application market just by dropping a quick beta into the iPhone Apps store.
That last bit about "reasonable patents and copyrights" says you still own your code, and Apple can't use it directly without licensing it. Sure, they can spend some of their own development resources writing their own version of a program if yours happens to become popular, but so can every other software house out there.
Point 1: Google's been collecting photos since 2006 and still hasn't completed a first pass.
Point 2: The odds of anyone seeing any specific image already on Google Streets are vanishingly small.
Point 3: Even if people appear in the photos, they aren't identified by name.
Point 4: The vast majority of places where a person can be photographed are neither incriminating nor embarrassing.
That's a far cry from Orwellian surveillance. As for deterrent effect, I could spend a week standing on the corner with a rocket launcher in one hand and a big sign that says, "I'm going to blow up a bank vault," in the other, and still have a better chance of winning the lottery than getting caught because someone caught me through Google Street.
The N-nines model is a fast and easy way to compare order-of-magnitude differences between existing networks, but it says almost nothing meaningful about actual usage or the perception of uptime from a user's perspective.
.. nobody cares if an incoming piece of email got delayed by 30 seconds at the MTU, but they do get testy if they can't load their webpages. But web surfing only uses 1-2 seconds of bandwidth per minute anyway.
.. who cares about a half hour of downtime from 0300 to 0330 when no one in your company is actually in the building and using the network?
Let's look at the numbers: 99.9% uptime translates to about 9 hours of unscheduled downtime a year. That can be one 9-hour block once a year, 36 minutes per day, 1.5 minutes per hour, 1.5 seconds per minute, or one dropped packet per thousand. Sure, it's easy to spot a 9-hour blackout, but as the slices of downtime get thinner, they get harder to notice at all, or to identify as USD specifically.
99.999% uptime translates to about 5 minutes of USD per year, and is of questionable value. You can't identify a network outage, call in a complaint, and get the issue resolved in the given timeframe. 99.9999%? It is to laugh. You can't even look up the tech support phone number without blowing your downtime budget for the year. Get hit by a rolling blackout for an hour? Kiss your downtime budget goodbye for the next 120 years.
Getting back to 99.9% uptime, let's move on to standard utilization patterns. USD really only becomes an issue if people notice it
If we have 2 seconds of usage and 2 seconds of downtime per minute, the odds of a collision are around 15:1 with an average overlap of 1 second when a collision does happen. Simply interleaving usage and downtime that way increases the perceived uptime by an order of magnitude since 90% of the outages happen when no one is actually using the network. And larger blocks of downtime get lost in larger blocks of non-utilization exactly the same way
Granted, if you have higher utilization you'll have a better chance of hitting a chunk of downtime, but you'll also have higher chances of queuing latency within your own use patterns. If you're already using 99% of your bandwidth, you can't just plunk in one more job and expect it to run immediately. It has to wait for that 1% of space no one else is currently using. And when you get to that point, it's really time to consider buying a bigger pipe anyway.
And that brings us to the main point: People don't buy network connectivity in absolute terms. They buy capacity, and the capacity they buy is scaled to what they think of as acceptable peak usage. "Acceptable peak usage" is a subjective thing, and nobody makes subjective judgements with 99.999% precision.