Of course, Apple is a hardware company. But they have no desire to be a commodity hardware company and compete purely on price/features. That's why they develop their own OS: the OS is likely a loss leader, but it (a) provides perceived added value that can only be used on their hardware (b) locks users in to paying for Apple's hardware if they want to keep using Mac-only software and file formats.
Likewise, iTunes is a loss leader (or maybe a breakeven product). But it encourages repeat purchases of Apple hardware by (a) forcing users to move to different software if they want to use a different player, and (b) forcing users to abandon any DRMed content that they have purchased through iTunes.
If you are very happy with a Lenovo ThinkPad computer or with a Toshiba DVD player, your reason for getting another one is to be sure that you get another one of similar quality, but there's nothing else really stopping you from buying a Dell, or an Sony, or a Samsung unit if there's a really good deal or a superior model. If for some reason you prefer a new Vaio or a Sandisk Sansa as a successor to a Mac or an iPod, there are a number of highly significant considerations you need to take into account that might make you reconsider changing manufacturers. This is one of the cornerstones of Apple's profit model, and if you think this hasn't occurred to Steve Jobs, you're crazy.
DRM isn't the only thing that helps ensure people will stay loyal to iPods, but nevertheless it's one of several important factors that Apple uses to try to lock people in. When you add in the fact that other DRM schemes such as PlaysForSure are so bar that they actively deter customers from buying products that use them, DRM is a massive competitive advantage. Why would Apple want to give that up?
Even though I disagree with some of your overall conclusions, in general, I think you've made an excellent, highly reasonable post here.
One point I would make:
And my point about the hurricane season was that if they can't even get the short term predictions correct, how can they get the long term ones correct?
Well, I hope you will excuse my use of an analogy here, but I think it's germane to the problem. If I'm running a retail store, it's perfectly natural for me to want to do sales forecasting. Now, this can be done in some kind of back of the envelope way such as last year's sales * 1.15 = this year's projected sales, or it can be done using an elaborate regression model that requires a team of economists to produce. Anyway, suppose you take extensive care and build a good model.
Now suppose you start trying to forecast what the sales will be for, say, February 23rd. Odds are, the daily data is drastically noisier than the annual data. While even a lousy annual model is quite likely to be accurate within +/-50% overall, it might be that for a certain type of store on a given day you sell anywhere between $500 and $10,000 worth of products, depending on who walks through your door and the value of the products they choose to purchase.
At a micro level, your model may well be effectively worthless. But it still may have significant value at making those larger scale predictions, because the low-level noise tends to balance out over time.
Obviously, this is nowhere near proof that any given climate model is valid. But it could well be that a climate model is no more accurate at predicting the severity of a given winter or hurricane season than The Old Farmer's Almanac, but it quite effective at predicting average climate trends over a longer period. So I don't believe that your assertion is valid that models must first prove themselves highly accurate at short term data before any weight should be applied to their prediction for long term data.
The main reason that you don't see many people surf the web from a phone is because the screen on a phone (or even on the largest Pocket PCs) is too small to comfortably display web pages. Unless Apple singlehandedly manages to convince every web designer that default web pages should be designed for iPhones first and computers second, the iPhone isn't going to change this. There's no evidence that the mobile version of Safari is any better than, say, the mobile version of Opera, which is available for various small-form-factor platforms already.
Besides, the iPhone isn't even going to have cellular broadband, which puts it a cut below various similar devices until you happen to be in range of a WiFi hotspot.
Necessarily liberty can flourish only amongst men of good will; conversely, it will always wither and die when granted to liars and thieves.
It's nice to see that we need never worry about trying to further democracy again. After all, if you have a population fully composed of just, good individuals, then surely one of them running a benevolent dictatorship is as good as having them all elect a government. And if there is any corruption, then democracy and freedom could never flourish anyway.
Alternatively, I suppose you could believe in a system of checks and balances, designed to allow a government to survive amongst an imperfect population ruled by imperfect representatives. But that would require moral complexity, which you don't seem to be comfortable with.
It's a start. My guess is that we'll see more of this, but it may take a couple of years.
It's been discussed that the point of the ATI acquisition is putting GPU cores onboard the CPU in future chips anyway, and outboard graphics card (and even discrete onboard chips) will start to decline in importance.
If that's true, you have to figure it will all end up open sourced eventually -- proprietary drivers for code that's onboard the CPU just aren't going to fly for very long.
The CPU is already designed, and has been in production for a while. (Corollary: that's why it's so cheap.)
It's an AMD Geode, which is an x86 computer-on-a-chip (onboard graphics, io, memory controller). It's a little late to try rearchitecting it to make it work differently.
Suppose we completely set aside the goal of widespread adoption, and simply leave open the goal that people should be able to use Linux to do whatever they want. That's not too controversial, right?
That means people should be able to do 3D design and games on Linux, at least ideally.
That means that either:
1. ATI and Nvidia should be forced to open up their drivers. Great idea, but no one has been able to do it.
2. Someone should reverse engineer 3D acceleration for graphics cards. Another great idea, but it's proven to be really hard so far.
3. Someone should start a business to produce 3D graphics cards with open source drivers. This is nice in theory, but it's a pretty difficult business to break into.
4. All 3D software should be designed so that no hardware acceleration is required. Nice in theory, doesn't work in the real world.
5. If a category of software can't be done without hardware acceleration, no one should ever use it. This one is obviously completely unrealistic.
6. Closed-source drivers are a fact of life for the foreseeable future.
7. If you want to use software that requires hardware 3D acceleration, you have to pick a closed source OS -- either OS X or Windows.
For now, the only realistic choice is 6 or 7. Of course, if you have the luxury of not needing 3D acceleration, you may not personally have to choose between these options. But other people do.
Let me mull over a couple of things you found to be most telling:
Also, while it may sound dumb to us, you know that they focus-tested the hell out of it in all three territories and, at the very least, it's not completely repellent to those focus groups.
Do I understand you correctly? Any branding that comes from a major corporation cannot suck by definition, because it has been tested with focus groups?
This is basically Nintendo trying to create a name and brand that is in no way similar to the others, in order to be distinct in the minds of consumers.
I think it's a lot more likely to make everyone who hears it say "WTF?", including the parents of small children that they want to buy this thing. I can't see how that helps the console at all.
Igor International or not, this makes sense to me:
"The biggest key to figuring out it's a bad name is when they explain it," he said. "You don't have to explain a good name, you have to explain a bad name."
The fact is that theoretical CS courses don't teach "fundamentals". They teach theoretical CS, which overlaps somewhat but not entirely with what people usually get paid to do in industry. People who think they're going to walk into their first job and spend all day implementing graph algorithms will get a nasty shock....
Oddly enough, I work with graph traversal a lot -- I use graphs extensively for managing relational data.
I also think a good understanding of algorithm design is among a number of tools that can help you build a shop with a modest number of really good programmers rather than a large number of iffy programmers -- and a shop full of low-value-added programmers is the kind of place that's perfect to be outsourced to India or China.
Of course, a degree doesn't make someone a really good programmer, and not having a piece of paper won't prevent you from becoming one. But a lot of the coursework in a CS degree can be really useful under the right circumstances.
Programming primarily consists of things like inventory, payroll, scheduling, POS devices, and things like that. Sure there are some places for CS people where performance is pushing the envelope like CAD/CAM/games, but those are a small percentage of what is being programed on a daily basis.
Summary of parent: my job can be done by monkeys or automated away entirely.
Even better -- install it into VMWare Player and try it out running on top of Windows. You can download ready-to-go images of Ubuntu and a bunch of other distros at VMWare and try them out without messing around with installation procedures, partitioning, or anything else.
Heck, if you install Xming or Cygwin/X, you can even run Linux apps inside your Windows environment until you're ready to cut over for good.
Having discrete devices that actually work together is a terrific idea. The problem is that carrying around multiple devices today leaves you with less than the sum of the parts, because the devices don't work with each other.
That's the reason I'm interested in a convergence device.
A high-end PDA phone with a big flash disk is a good enough phone, music & video player, PDA, game player, etc. to make up for the fact that it isn't best-in-class at any of these things. But only because the separate devices don't work together, so carrying around a "utility belt" full of different devices is utterly worthless.
Sure, but it's not a huge percentage of PC sales either here or there. If you buy parts, the OS is your problem, just like always.
But if you buy your PC at Best Buy (or its Chinese equivalent), it probably ought to come with an OS anyway, and that's all that's being accomplished. If it's Linux, so much the better.
They might release a UMD burner drive, but it's sure not going to be sold as a way to copy DVD movies to the PSP. Sony is/owns a movie studio. As far as the movie studios are concerned, under the DMCA, format-shifting DVDs that you have legally purchased is piracy.
I bought a $125 mower, because it had the features I needed.
Actually, I also bought a $125 mower, and it has all the features I need. It was made in the USA, and it's low maintenance, very high quality, and is perfectly fuel-efficient.
Of course, it doesn't have a motor at all. It's a reel mower.
Exactly -- if you need to embed a browser in your application, you can't just embed "default browser."
And there may be design constraints compelling IE over an alternate component -- we embedded the MSHTML component in our application because the primary use is to access MS Reporting Services which renders horribly in other browsers (but that's a different topic I won't go into further).
You can still encounter weirdness in other browsers even if it works in Firefox and IE. Opera and Safari are the most important other browsers to check -- along with IE 5, which still shows up on the browser usage charts.
Depending on your target audience, you may want to test other browsers as well, such as Konqueror (Linux), NetFront (handhelds), or v4 IE & NN.
But the technology that made it so easy for her to do it was: internet enabled e-mail.
It's pretty easy to tamper with witnesses and perform all sorts of other stupid or illegal communications over telephones. If we don't trust federal agents to use communications technologies properly, maybe they shouldn't have phones, either.
My point is that the "cost" of turning on publicly-transcieving e-mail accounts for investigators and other people with legally critical jobs involve more than some server admin mouseclicks and a little more storage... there's substantial training and oversight involved.
Lawyers and stockbrokers have managed to get past this hurdle. I would think that highly trained FBI agents could handle it as well.
It is unfortunate that it ties the client presentation layer with a MS proprietary web server...but it is Microsoft after all..
It'd be nice if the two could be independent
It's not any different than the nifty Ajax accelerators that ship with Rails -- the whole point is that it's coupled with the server development platform, because the idea is to let you seamlessly use one development platform for both the client and server work.
If you want server-independent client-side JS libraries, there are several good ones, such as scriptaculous. This is something different.
I'm no MS fanboy, but since when is coining a term the same thing as inventing a technology?
The XMLHttpRequest object -- which is the technical basis for AJAX -- was developed by Microsoft years before the name AJAX was coined. The style of coding that uses asynchronous javascript requests to return XML data for live updates instead of page refreshes was first used in the web version of Outlook.
The object was originally developed by Microsoft Corp., available since Internet Explorer 5.0 as an ActiveX object, via JScript, VBScript, or other scripting languages supported by the browser. Mozilla contributors then implemented a compatible native version in Mozilla 1.0. This was later followed by Apple since Safari 1.2, Opera Software since Opera 8.0 and iCab since 3.0b352. Additionally, Open Laszlo added support in version 3.1.
Slashdot Mirror
It's a database with full-on ACID transactions in 250K of SQL. I laugh in your face, Oracle.
Having used openoffice, I've made the switch to google docs. I get 80% of the functions with 20% of the hassle.
I think you mean 20% of the functionality. I am frankly astounded by how few features Google Docs has.
Having said that, if it works for you, more power to you.
like all of Apple's recent products, it will 'Just Work'.
You mean like the ROKR? Apple fans are always quick to disavow that one as though Apple had never touched it.
It will be a hybrid iPod/cell phone/PDA with no sacrifices in functionality, compared to carrying around three separate devices.
Wouldn't a lack of 3G be a sacrifice in functionality?
As Jobs mentioned in his keynote, the price is still cheaper than buying a smartphone and iPod Nano separately.
Well, you can buy a Pocket PC phone for $2-300 and drop in an SD card to hold music and movies, so I'm not sure what a Nano would bring to the party.
Of course, Apple is a hardware company. But they have no desire to be a commodity hardware company and compete purely on price/features. That's why they develop their own OS: the OS is likely a loss leader, but it (a) provides perceived added value that can only be used on their hardware (b) locks users in to paying for Apple's hardware if they want to keep using Mac-only software and file formats.
Likewise, iTunes is a loss leader (or maybe a breakeven product). But it encourages repeat purchases of Apple hardware by (a) forcing users to move to different software if they want to use a different player, and (b) forcing users to abandon any DRMed content that they have purchased through iTunes.
If you are very happy with a Lenovo ThinkPad computer or with a Toshiba DVD player, your reason for getting another one is to be sure that you get another one of similar quality, but there's nothing else really stopping you from buying a Dell, or an Sony, or a Samsung unit if there's a really good deal or a superior model. If for some reason you prefer a new Vaio or a Sandisk Sansa as a successor to a Mac or an iPod, there are a number of highly significant considerations you need to take into account that might make you reconsider changing manufacturers. This is one of the cornerstones of Apple's profit model, and if you think this hasn't occurred to Steve Jobs, you're crazy.
DRM isn't the only thing that helps ensure people will stay loyal to iPods, but nevertheless it's one of several important factors that Apple uses to try to lock people in. When you add in the fact that other DRM schemes such as PlaysForSure are so bar that they actively deter customers from buying products that use them, DRM is a massive competitive advantage. Why would Apple want to give that up?
Even though I disagree with some of your overall conclusions, in general, I think you've made an excellent, highly reasonable post here.
One point I would make:
And my point about the hurricane season was that if they can't even get the short term predictions correct, how can they get the long term ones correct?
Well, I hope you will excuse my use of an analogy here, but I think it's germane to the problem. If I'm running a retail store, it's perfectly natural for me to want to do sales forecasting. Now, this can be done in some kind of back of the envelope way such as last year's sales * 1.15 = this year's projected sales, or it can be done using an elaborate regression model that requires a team of economists to produce. Anyway, suppose you take extensive care and build a good model.
Now suppose you start trying to forecast what the sales will be for, say, February 23rd. Odds are, the daily data is drastically noisier than the annual data. While even a lousy annual model is quite likely to be accurate within +/-50% overall, it might be that for a certain type of store on a given day you sell anywhere between $500 and $10,000 worth of products, depending on who walks through your door and the value of the products they choose to purchase.
At a micro level, your model may well be effectively worthless. But it still may have significant value at making those larger scale predictions, because the low-level noise tends to balance out over time.
Obviously, this is nowhere near proof that any given climate model is valid. But it could well be that a climate model is no more accurate at predicting the severity of a given winter or hurricane season than The Old Farmer's Almanac, but it quite effective at predicting average climate trends over a longer period. So I don't believe that your assertion is valid that models must first prove themselves highly accurate at short term data before any weight should be applied to their prediction for long term data.
The main reason that you don't see many people surf the web from a phone is because the screen on a phone (or even on the largest Pocket PCs) is too small to comfortably display web pages. Unless Apple singlehandedly manages to convince every web designer that default web pages should be designed for iPhones first and computers second, the iPhone isn't going to change this. There's no evidence that the mobile version of Safari is any better than, say, the mobile version of Opera, which is available for various small-form-factor platforms already.
Besides, the iPhone isn't even going to have cellular broadband, which puts it a cut below various similar devices until you happen to be in range of a WiFi hotspot.
Necessarily liberty can flourish only amongst men of good will; conversely, it will always wither and die when granted to liars and thieves.
It's nice to see that we need never worry about trying to further democracy again. After all, if you have a population fully composed of just, good individuals, then surely one of them running a benevolent dictatorship is as good as having them all elect a government. And if there is any corruption, then democracy and freedom could never flourish anyway.
Alternatively, I suppose you could believe in a system of checks and balances, designed to allow a government to survive amongst an imperfect population ruled by imperfect representatives. But that would require moral complexity, which you don't seem to be comfortable with.
It's a start. My guess is that we'll see more of this, but it may take a couple of years.
It's been discussed that the point of the ATI acquisition is putting GPU cores onboard the CPU in future chips anyway, and outboard graphics card (and even discrete onboard chips) will start to decline in importance.
If that's true, you have to figure it will all end up open sourced eventually -- proprietary drivers for code that's onboard the CPU just aren't going to fly for very long.
The CPU is already designed, and has been in production for a while. (Corollary: that's why it's so cheap.)
It's an AMD Geode, which is an x86 computer-on-a-chip (onboard graphics, io, memory controller). It's a little late to try rearchitecting it to make it work differently.
Suppose we completely set aside the goal of widespread adoption, and simply leave open the goal that people should be able to use Linux to do whatever they want. That's not too controversial, right?
That means people should be able to do 3D design and games on Linux, at least ideally.
That means that either:
1. ATI and Nvidia should be forced to open up their drivers. Great idea, but no one has been able to do it.
2. Someone should reverse engineer 3D acceleration for graphics cards. Another great idea, but it's proven to be really hard so far.
3. Someone should start a business to produce 3D graphics cards with open source drivers. This is nice in theory, but it's a pretty difficult business to break into.
4. All 3D software should be designed so that no hardware acceleration is required. Nice in theory, doesn't work in the real world.
5. If a category of software can't be done without hardware acceleration, no one should ever use it. This one is obviously completely unrealistic.
6. Closed-source drivers are a fact of life for the foreseeable future.
7. If you want to use software that requires hardware 3D acceleration, you have to pick a closed source OS -- either OS X or Windows.
For now, the only realistic choice is 6 or 7. Of course, if you have the luxury of not needing 3D acceleration, you may not personally have to choose between these options. But other people do.
Let me mull over a couple of things you found to be most telling:
Also, while it may sound dumb to us, you know that they focus-tested the hell out of it in all three territories and, at the very least, it's not completely repellent to those focus groups.
Do I understand you correctly? Any branding that comes from a major corporation cannot suck by definition, because it has been tested with focus groups?
This is basically Nintendo trying to create a name and brand that is in no way similar to the others, in order to be distinct in the minds of consumers.
I think it's a lot more likely to make everyone who hears it say "WTF?", including the parents of small children that they want to buy this thing. I can't see how that helps the console at all.
Igor International or not, this makes sense to me:
"The biggest key to figuring out it's a bad name is when they explain it," he said. "You don't have to explain a good name, you have to explain a bad name."
The fact is that theoretical CS courses don't teach "fundamentals". They teach theoretical CS, which overlaps somewhat but not entirely with what people usually get paid to do in industry. People who think they're going to walk into their first job and spend all day implementing graph algorithms will get a nasty shock....
Oddly enough, I work with graph traversal a lot -- I use graphs extensively for managing relational data.
I also think a good understanding of algorithm design is among a number of tools that can help you build a shop with a modest number of really good programmers rather than a large number of iffy programmers -- and a shop full of low-value-added programmers is the kind of place that's perfect to be outsourced to India or China.
Of course, a degree doesn't make someone a really good programmer, and not having a piece of paper won't prevent you from becoming one. But a lot of the coursework in a CS degree can be really useful under the right circumstances.
Programming primarily consists of things like inventory, payroll, scheduling, POS devices, and things like that. Sure there are some places for CS people where performance is pushing the envelope like CAD/CAM/games, but those are a small percentage of what is being programed on a daily basis.
Summary of parent: my job can be done by monkeys or automated away entirely.
Or did I miss something?
Even better -- install it into VMWare Player and try it out running on top of Windows. You can download ready-to-go images of Ubuntu and a bunch of other distros at VMWare and try them out without messing around with installation procedures, partitioning, or anything else.
Heck, if you install Xming or Cygwin/X, you can even run Linux apps inside your Windows environment until you're ready to cut over for good.
Having discrete devices that actually work together is a terrific idea. The problem is that carrying around multiple devices today leaves you with less than the sum of the parts, because the devices don't work with each other.
That's the reason I'm interested in a convergence device.
A high-end PDA phone with a big flash disk is a good enough phone, music & video player, PDA, game player, etc. to make up for the fact that it isn't best-in-class at any of these things. But only because the separate devices don't work together, so carrying around a "utility belt" full of different devices is utterly worthless.
Sure, but it's not a huge percentage of PC sales either here or there. If you buy parts, the OS is your problem, just like always.
But if you buy your PC at Best Buy (or its Chinese equivalent), it probably ought to come with an OS anyway, and that's all that's being accomplished. If it's Linux, so much the better.
They might release a UMD burner drive, but it's sure not going to be sold as a way to copy DVD movies to the PSP. Sony is/owns a movie studio. As far as the movie studios are concerned, under the DMCA, format-shifting DVDs that you have legally purchased is piracy.
I highly doubt MS is going to shun their own proprietary technology
They're already making the salespitch for their new, improved proprietary technology.
Microsoft EOLs old proprietary technologies to replace them with new flavors all the time. Just look at their history of database access technologies.
If you govern your life based on what will impress the neighbors, I feel sorry for you.
I bought a $125 mower, because it had the features I needed.
Actually, I also bought a $125 mower, and it has all the features I need. It was made in the USA, and it's low maintenance, very high quality, and is perfectly fuel-efficient.
Of course, it doesn't have a motor at all. It's a reel mower.
Exactly -- if you need to embed a browser in your application, you can't just embed "default browser."
And there may be design constraints compelling IE over an alternate component -- we embedded the MSHTML component in our application because the primary use is to access MS Reporting Services which renders horribly in other browsers (but that's a different topic I won't go into further).
You can still encounter weirdness in other browsers even if it works in Firefox and IE. Opera and Safari are the most important other browsers to check -- along with IE 5, which still shows up on the browser usage charts.
Depending on your target audience, you may want to test other browsers as well, such as Konqueror (Linux), NetFront (handhelds), or v4 IE & NN.
But the technology that made it so easy for her to do it was: internet enabled e-mail.
It's pretty easy to tamper with witnesses and perform all sorts of other stupid or illegal communications over telephones. If we don't trust federal agents to use communications technologies properly, maybe they shouldn't have phones, either.
My point is that the "cost" of turning on publicly-transcieving e-mail accounts for investigators and other people with legally critical jobs involve more than some server admin mouseclicks and a little more storage... there's substantial training and oversight involved.
Lawyers and stockbrokers have managed to get past this hurdle. I would think that highly trained FBI agents could handle it as well.
It is unfortunate that it ties the client presentation layer with a MS proprietary web server...but it is Microsoft after all..
It'd be nice if the two could be independent
It's not any different than the nifty Ajax accelerators that ship with Rails -- the whole point is that it's coupled with the server development platform, because the idea is to let you seamlessly use one development platform for both the client and server work.
If you want server-independent client-side JS libraries, there are several good ones, such as scriptaculous. This is something different.
I'm no MS fanboy, but since when is coining a term the same thing as inventing a technology?
The XMLHttpRequest object -- which is the technical basis for AJAX -- was developed by Microsoft years before the name AJAX was coined. The style of coding that uses asynchronous javascript requests to return XML data for live updates instead of page refreshes was first used in the web version of Outlook.
From Wikipedia:
The object was originally developed by Microsoft Corp., available since Internet Explorer 5.0 as an ActiveX object, via JScript, VBScript, or other scripting languages supported by the browser. Mozilla contributors then implemented a compatible native version in Mozilla 1.0. This was later followed by Apple since Safari 1.2, Opera Software since Opera 8.0 and iCab since 3.0b352. Additionally, Open Laszlo added support in version 3.1.