Some thoughts & history on OS mass adoption
There are constant debates on the ubuntuforums about the "readiness" of Linux for mass adoption. People contend that in order for the masses to move to Linux, certain things must change--their webcam has to "just work," or Linux should run iTunes, or the user interface has to behave exactly like Windows, and so on.
I have consistently argued the case that Microsoft's market dominance has NEVER depended on constant improvement, and has always depended on more nitty-gritty economic factors. Geeks--especially Free Software geeks!--hate thinking about economics and history for reasons I can't understand. I can only speculate that it has to do with their inherent allergy to money and markets in general.
There are other things--and by that I mean other than the software itself-- you have to think about when it comes to mass adoption.
* TIME TO MARKET/ MARKET PENETRATION. The personal computer era as we know it really began when IBM released the Model 5120 PC in 1981. The 'superior' systems did not release until later--The Apple Macintosh was released in 1984, Commodore's Amiga in 1986. That gave IBM, Microsoft, Intel, and the clonemakers a staggering three to five year head-start on their competition.
* COST. Think back to the dawn of personal computing. How much would you have paid for a computer and operating system? With a little sleight-of-economic-hand, we can recreate the buying decision in the year 1986: (note 1)
-- IBM PERSONAL COMPUTER, MODEL 5120, with MS-DOS: $1565 (1981 US Dollars). Inflation-adjusted for 1986 dollars: $2044 . Inflation adjusted for 2005 (note 2)? $3502 !
-- APPLE MACINTOSH (with its elegant, beautiful interface): $2495 US dollars in 1984. Inflation-adjusted 2005 dollars: $4618.
-- COMMODORE AMIGA 1000, $1500 in 1986 (!) 2005 dollars: $2570.
* USABILITY. This is a no-brainer. The masses hate command lines. The masses love GUIs, especially pretty GUIs. Cost no object, it's Apple and Amiga for the win.
* PERFORMANCE. This is a little more complicated. As much as we complain about how creaky MS-DOS is/was as an operating system, I don't remember it being so bad at the time. DOS apps for the IBM PC were lean, mean, and surprisingly fast, considering the hardware on which they had to run. Laboriously hand-coded in x86 assembly, WordPerfect was THE program to beat for word processing in the '80s. Not because it was pretty (it wasn't) or because it was particularly easy to use (Three or more levels of functionality for every function key, thanks to function-key-chording. Somewhere we still have the F-key overlay that told us which F-key did what). WordPerfect won because it was FAST and EFFICIENT. Ditto Lotus 1-2-3.
Meanwhile, there were the various problems associated with implementing preemptive multitasking on the GUIfied systems. Who can forget the Macintosh's cute (but ominous) "bomb" error, or Amiga's "Guru meditation"? Let's face it--none of the PCs of the '80s had great memory management.
Now let's put this all together.
By 1986, it was possible to move to a 'superior' platform--the Amiga. It was technically superior to the IBM, and cost about the same. It was miles ahead in usability. Why did the future not belong to Commodore?
Well, by 1986, there were already hundreds of thousands of IBM-compatible machines. Remember, IBM and MS-DOS were on the scene **first** and gained all of the first-mover advantages.
It's not as if there was no competition--there was. Not from the Macintosh, of course--one look at the price should tell you why. Competition came instead from The Clones, the "IBM-compatibles" which drove down retail prices for hardware. So the "list" price on the Model 5120 can be treated as an upper bound; there were MS-DOS machines on the market that were significantly cheaper. With hardware prices dropping, and a proven market for software, the procurement types in large organizations did what they had to do--they bought IBM, IBM-compatible, and Microsoft. "Nobody was ever fired for buying from IBM," after all. The initial cost savings over Amigas and Apples were significant, and the efficiency/reliability of the relatively unsophisticated single-user, single-task design won out over the elegance (and instability) of GUIs and preemptive multitasking.
Note that the computer was essentially a *business* machine, whose arrival at home also signified the ability to take work home from the office. People bought IBM-compatibles because that's what they were using at work--not because it was a blatantly superior system for home use.
So the adoption loop, as you see, has very little to do with the sort of chrome that people like to think about. Usability, features, and polish were less important than the less glamorous world of economics--efficiency, return-on-investment, and plain old price.
Where does this leave Linux? In a very good place, surprisingly enough. The tech industry has already begun a rapid race to the bottom, at least as far as costs go. Free and Open-source software's single greatest economic virtue is its low development costs: there is no need to reinvent the wheel when you have the right to build upon previous implementations. Community support factors in here as well--because the community has the right to use and build upon your software, they can contribute and reap immediate benefits.
Does this mean that developer pay will fall? Probably. Does it also mean that the cost of all software, for all users, is also likely to fall? Almost certainly. In the end, it is this economic advantage that will slowly drive the adoption of Free Software.
Proprietary vendors like Microsoft, Apple, and Adobe will fight valiant rearguard actions, to be sure. I predict a veritable free-fall in the Windows and Office license fees for large organizations, almost to the point that software is being offered gratis.
But what about freedom? There's the rub: when the cost of all software has fallen to zero, or close to zero, competition shifts away from cost to other things. If all software is gratis, then why wouldn't you also get the software that does what you want? Or the software that enables you to run it in the way that you want it, without fear of legal liability?
The Free Software Revolution, like the PC revolution that made it possible, is no tea party: it will come about because of the economic self-interest of people who control LARGE NETWORKS of computers, because millions of home users were "enlightened" and began to heed the words of RMS and the Free Software Foundation. And if it doesn't happen before the next Ubuntu release cycle, or the next Kernel version, or the final release of GNU/HURD, that still doesn't change the fundamental cost advantages available to free software projects. It's a long way down, folks--enjoy the view from here, and enjoy the ride.
What I want to ask is--why are economic arguments not avanced enough in the Free Software community? For all the dozens of F/LOSS Thomas Jefferson types, why are there so few Alexander Hamiltons who can make the synthesis of freedom and markets, of self-interest and commonwealth?
------
1. Price sources: IBM Price for Model 5120, IBM ARchives, Wikipedia on the Macintosh, for the Amiga: http://www.old-computers.com/museum/computer.asp?c=28
2. Inflation/CPI calculator All 'current' prices indexed to 2005, the last year of available CPI data.
I have consistently argued the case that Microsoft's market dominance has NEVER depended on constant improvement, and has always depended on more nitty-gritty economic factors. Geeks--especially Free Software geeks!--hate thinking about economics and history for reasons I can't understand. I can only speculate that it has to do with their inherent allergy to money and markets in general.
There are other things--and by that I mean other than the software itself-- you have to think about when it comes to mass adoption.
* TIME TO MARKET/ MARKET PENETRATION. The personal computer era as we know it really began when IBM released the Model 5120 PC in 1981. The 'superior' systems did not release until later--The Apple Macintosh was released in 1984, Commodore's Amiga in 1986. That gave IBM, Microsoft, Intel, and the clonemakers a staggering three to five year head-start on their competition.
* COST. Think back to the dawn of personal computing. How much would you have paid for a computer and operating system? With a little sleight-of-economic-hand, we can recreate the buying decision in the year 1986: (note 1)
-- IBM PERSONAL COMPUTER, MODEL 5120, with MS-DOS: $1565 (1981 US Dollars). Inflation-adjusted for 1986 dollars: $2044 . Inflation adjusted for 2005 (note 2)? $3502 !
-- APPLE MACINTOSH (with its elegant, beautiful interface): $2495 US dollars in 1984. Inflation-adjusted 2005 dollars: $4618.
-- COMMODORE AMIGA 1000, $1500 in 1986 (!) 2005 dollars: $2570.
* USABILITY. This is a no-brainer. The masses hate command lines. The masses love GUIs, especially pretty GUIs. Cost no object, it's Apple and Amiga for the win.
* PERFORMANCE. This is a little more complicated. As much as we complain about how creaky MS-DOS is/was as an operating system, I don't remember it being so bad at the time. DOS apps for the IBM PC were lean, mean, and surprisingly fast, considering the hardware on which they had to run. Laboriously hand-coded in x86 assembly, WordPerfect was THE program to beat for word processing in the '80s. Not because it was pretty (it wasn't) or because it was particularly easy to use (Three or more levels of functionality for every function key, thanks to function-key-chording. Somewhere we still have the F-key overlay that told us which F-key did what). WordPerfect won because it was FAST and EFFICIENT. Ditto Lotus 1-2-3.
Meanwhile, there were the various problems associated with implementing preemptive multitasking on the GUIfied systems. Who can forget the Macintosh's cute (but ominous) "bomb" error, or Amiga's "Guru meditation"? Let's face it--none of the PCs of the '80s had great memory management.
Now let's put this all together.
By 1986, it was possible to move to a 'superior' platform--the Amiga. It was technically superior to the IBM, and cost about the same. It was miles ahead in usability. Why did the future not belong to Commodore?
Well, by 1986, there were already hundreds of thousands of IBM-compatible machines. Remember, IBM and MS-DOS were on the scene **first** and gained all of the first-mover advantages.
It's not as if there was no competition--there was. Not from the Macintosh, of course--one look at the price should tell you why. Competition came instead from The Clones, the "IBM-compatibles" which drove down retail prices for hardware. So the "list" price on the Model 5120 can be treated as an upper bound; there were MS-DOS machines on the market that were significantly cheaper. With hardware prices dropping, and a proven market for software, the procurement types in large organizations did what they had to do--they bought IBM, IBM-compatible, and Microsoft. "Nobody was ever fired for buying from IBM," after all. The initial cost savings over Amigas and Apples were significant, and the efficiency/reliability of the relatively unsophisticated single-user, single-task design won out over the elegance (and instability) of GUIs and preemptive multitasking.
Note that the computer was essentially a *business* machine, whose arrival at home also signified the ability to take work home from the office. People bought IBM-compatibles because that's what they were using at work--not because it was a blatantly superior system for home use.
So the adoption loop, as you see, has very little to do with the sort of chrome that people like to think about. Usability, features, and polish were less important than the less glamorous world of economics--efficiency, return-on-investment, and plain old price.
Where does this leave Linux? In a very good place, surprisingly enough. The tech industry has already begun a rapid race to the bottom, at least as far as costs go. Free and Open-source software's single greatest economic virtue is its low development costs: there is no need to reinvent the wheel when you have the right to build upon previous implementations. Community support factors in here as well--because the community has the right to use and build upon your software, they can contribute and reap immediate benefits.
Does this mean that developer pay will fall? Probably. Does it also mean that the cost of all software, for all users, is also likely to fall? Almost certainly. In the end, it is this economic advantage that will slowly drive the adoption of Free Software.
Proprietary vendors like Microsoft, Apple, and Adobe will fight valiant rearguard actions, to be sure. I predict a veritable free-fall in the Windows and Office license fees for large organizations, almost to the point that software is being offered gratis.
But what about freedom? There's the rub: when the cost of all software has fallen to zero, or close to zero, competition shifts away from cost to other things. If all software is gratis, then why wouldn't you also get the software that does what you want? Or the software that enables you to run it in the way that you want it, without fear of legal liability?
The Free Software Revolution, like the PC revolution that made it possible, is no tea party: it will come about because of the economic self-interest of people who control LARGE NETWORKS of computers, because millions of home users were "enlightened" and began to heed the words of RMS and the Free Software Foundation. And if it doesn't happen before the next Ubuntu release cycle, or the next Kernel version, or the final release of GNU/HURD, that still doesn't change the fundamental cost advantages available to free software projects. It's a long way down, folks--enjoy the view from here, and enjoy the ride.
What I want to ask is--why are economic arguments not avanced enough in the Free Software community? For all the dozens of F/LOSS Thomas Jefferson types, why are there so few Alexander Hamiltons who can make the synthesis of freedom and markets, of self-interest and commonwealth?
------
1. Price sources: IBM Price for Model 5120, IBM ARchives, Wikipedia on the Macintosh, for the Amiga: http://www.old-computers.com/museum/computer.asp?c=28
2. Inflation/CPI calculator All 'current' prices indexed to 2005, the last year of available CPI data.