Matt Yglesias asks
Even at the height of Microsoft’s power in the late-’90s, Windows 98 was oddly ugly. Surely the richest company on the planet could hire someone to design a better logo than this, right? Why were the default color combinations on Excel charts so wretched? Why didn’t anyone else bother to design power adaptors that look good?
The story I was telling well into the mid 2000s was that Steve Jobs thought computers should be easy to use and appealing to the eye. Bill Gates thought they should be cheaper than dirt. And, Bill Gates was right.
I think Felix Salmon is right about what really happened. The Gates strategy of a Model T in every garage was correct when the tech was new an expensive. You cut corners in every way possible, you do anything to push the price down. Send it out half baked, full of holes and ugly as hell, but just send it out cheap.
I think Bill Gates is still mostly right about software. Cheap rulz.
However, the price of hardware collapsed so fast that Jobs-esque hardware could be produced at prices people could easily afford. You could cut $40 off the price of the Ipad by just settling for something a bit uglier, but that wouldn’t really improve sales.
That’s also, why Apple has only been successful when the software was married to the hardware. To my knowledge no one is buying OS X to install on their own Linux box.
12 comments
Comments feed for this article
Friday ~ October 7th, 2011 at 2:11 pm
Yera Meyahu
That’s because they’re not selling it to install on your LInux box.
Still, the ‘hackintosh’ concept does exist, with real-world users living in that legal gray area.
Friday ~ October 7th, 2011 at 2:20 pm
db
“To my knowledge no one is buying OS X to install on their own Linux box.”
A company called Psystar was doing this but Apple shut them down.
I’d also note that Apple’s very limited product line and the resulting economies of scale makes it very hard to get a comprable PC laptop at the Apple pricepoint.
Friday ~ October 7th, 2011 at 2:29 pm
CJ
But Karl’s point is (part of) why Apple doesn’t let people install OS X on their generic boxes.
Friday ~ October 7th, 2011 at 2:35 pm
Winston
A ton of people install OS X on PC hardware but it is challenging because Apple places technical and legal hurdles in your way. This being said, I can think of several I know who are running OS X on unauthorized hardware. Were it legal and supported by Apple, a lot of folks would be doing so. I would gladly pay $250-$300 for OS X, but I’m sadly stuck with windows on my non-Apple desktop.
Friday ~ October 7th, 2011 at 2:36 pm
PrometheeFeu (@PrometheeFeu)
Also, if you install OSX on it, it’s not a “linux box” anymore.
Friday ~ October 7th, 2011 at 5:50 pm
Edwin Perello
Please remember Apple was never successful as a desktop/laptop manufacturer. Before the iPod and iPhone introduced the world at large to Apple, the company had a market share of… 1-3%. Now Apple holds ~10% desktop/laptop market share, but mostly because of the iPhone.
And compared to common Windows boxes, the prices of Apple desktops/laptops were much higher until recently. I would go as far as arguing the lower price is subsidized by their mobile lines.
Sunday ~ October 9th, 2011 at 11:55 am
Curt Doolittle
True. Although there is a lot of historical evidence to suggest that the tipping point is only somewhere in the 15%-20% range.
Friday ~ October 7th, 2011 at 8:33 pm
Curt Doolittle
Yes, that’s right. Hardware costs have collapsed.
The advantage Apple has is that it can control quality because they control the hardware. Microsoft can’t. Because microsoft can’t interfere in the competition between their OEM’s. And it’s from OEM’s that microsoft makes it’s money.
Now that Jobs is gone, Apple will come out in the next three years with a low priced desktop model set that uses plain old VGA connectors. WHen that happens, the entire windows based business will collapse over five to six years.
At that point Microsoft will have to either a) buy a hardware company and compete against apple, or b) simply die off as did Digital Equipment, or c) become a services company.
There is a pretty significant difference between software architectures of past and present. Microsoft technology was written in an era when hardware was slow. It uses “Strongly Typed” languages, and runs ‘processes’. Compare that to this web site, in which the entire language, and the entire content management system, are loaded and parsed whenever a page is requested. Something that woud have been simply impossible in the late 1980′s and early 1990′s. Softly Typed langages mean that it’s a lot less expensive for ‘average’ people to write code quickly. Transactional systems are possible because hardware, memory and disks are fast. Web content is less expensive because bandwidth is very high. So, what we’re seeing is the aging of not just companies, but the entire legacy code base on which their revenues (rents) are based.
Most technology innovation today is actually being done by non proprietary companies: Facebook and Google. Oracle is on the way to becoming Sun. Microsoft on the way to becoming a low rent IBM. On the web, tech is being written in softly typed languages (php and javascript) on transactional requests, rather than running processes.
Microsoft is a classic victim of the innovator’s dilemma. They cannot save themselves because they are too dependent upon their OEM channels. They woud get into legal trouble if they tried to get into the hardware business. And tehy lack the management to do so. As a rent-seeking organization consisting largely of rent-seekers, they would have a harder time competing on the market if they did buy their way into hardware. THey won’t make a windows-lite. Because if they did, they would allow companies to make the decision to not only transition, but consider abandoning windows altogether.
Saturday ~ October 8th, 2011 at 5:39 am
Negi
There’s nothing stopping you from writing in a softly typed language in windows.. The .net framework is actually built to do things like that and allow you to use common APIs between languages.
In fact, conversely, Apple has attached all of its iGear to the success of the “Objective C” language, which they have, stupidly, refused to write a garbage collection routine for.
Please stop writing things that you read in popsci magazines and stop being a tool, “Curt Doolittle”.
Sunday ~ October 9th, 2011 at 11:53 am
Curt Doolittle
(Um.. Since I built the largest privately held MSFT based consultancy in the US, and since I just acquired one of the most complex MSFT product technologies in the market I think I have an opinion based upon experience and necessity, not popular science. It’s just one of the companies I’ve built, but its relevant to the discussion.)
My comment is based upon the practical reality of trying to make MSFT technologies compete today against development cycles that are less expensive in the arena where most technological development is happening: consumer and social. And where the hardware cost of vertical-scaling (in process – windows) servers is higher than the hardware cost of horizontally scaling (transactional) servers.
(Was’t it Andresson who just said that Oracle is a declining irrelevant company? It’s becoming a Sun?)
The data is the data. The data says that evolutionary development using softly typed languages on open source software reduces all costs. It says that the php/apache model (transactional) is now more economically viable that the compiled model (process) because a) there are no licensing costs, b) hardware is effectively free, c) developers with less ‘talent’ can be employed on interface development, and d) the minority of performance demanding services can be then written in C/C++ and added to the package when needed by a very small highly talented group. Furthermore the transaction model scales horizontally better on cheap servers and virtual machines than does the process model.
Microsoft grew very rapidly because of Win3.1, because it included free networking, because the ease of use of Visual basic 3 AND because of cheap hardware. That’s because there were 300K c++ programmers and 7M basic programmers. Technology growth is largely dependent upon what the generation just out of college is using, and how many of them are using it. That’s where innovation occurs. In that set. The world is using softly typed languages because the web is where the money is, and because established companies have drastically reduced internal R&D, and moved almost entirely to packages.
(You may not remember that when Java/Javascript came out, MSFT management fell behind the curve at that time as well, and that’s when they came up with the .NET strategy.)
Karl is making a point about cheap hardware. I’m making the additional point that the INSTITUTIONS that are currently ORGANIZED to exploit previous economic conditions (expensive hardware) are so significantly entrenched in their existing model that the INNOVATORS DILEMMA will prevent them from taking the risk necessary to evolve their platform to meet the current market conditions.
I”m making that point because I’m more interested in our institutions both private and public over the long term than I am our short term economy — which is why I argue a different position than Karl does. We all have our messages, and I’m just “staying on message”. Karl is part of a wing of people that assume away the difficulty of innovating under most institutional frameworks. (He would disagree.) I’m part of the classical liberal wing that argues that our historical institutional frameworks were innovative and not efficient, and that if we seek to make them efficient they will not be innovative and if they are not innovative then we are just selling off the continent to immigrants – rent seekers. And nothing more.
Microsoft is in a difficult position because while they have the best development tool (Visual studio), and an entrenched talent base, they have not adopted the php language (because they can’t control it) despite the fact that providing debugging would be sufficiently advantageous to developers that they would use it. Microsoft is also in a difficult position with product quality because they don’t make the hardware, and so third parties control the perception of the quality of their products. To make matters worse, their code must be extremely elaborate and fault tolerant to account for that lack of control over hardware. (For example: Flash causes an inordinate number of OSX crashes. Drivers cause an inordinate number of windows crashes.)
Their strategy (software only) was developed in an era where it was high risk and high cost to get into the hardware business, and where they could mobilize many other people in many industries to take hardware risks to drive down costs for consumers in order to get them cheap computers. (In other words they used other people’s money to drive costs down.) But today we are in a world where hardware is all but free – it’s certainly commodity priced – and consumers want to buy ‘appliances’ not ‘tools’. THe microsoft stack is designed as a tool. Apple stack is designed as an appliance.
Like I said. I am pretty sure that post-Jobs, Apple will come out with a desktop replacement line that doesn’t include a monitor, and that when they do, it will rapidly shift the economics of current technology. I think they may do that just because it will be difficult to envision a future in a company that has lost it’s visionary leader. Most companies falter when the visionary leader departs. So they seek visible rather than visionary opportunities.
I think until that happens, the MSFT management and board will continue their mining of the windows “Network Effect” because it sustains the share price. But, the company responds well to threats. It will likely change when threatened, and only at that point. Because at that point the threat to their share price will be higher than the promise of the network effect of the entrenched revenue streams. And because the government will not be able to INTERFERE and call them anti-competitive when they acquire HP’s computer division or Dell or someone of that nature.
Consumers suffer with more expensive goods and with a delay in obtaining new products because of a) STICKINESS of INSTITUTIONS, b) INNOVATOR’S DILEMMA, and c) GOVERNMENT INTERFERENCE with any significant action microsoft might take.
Wednesday ~ October 12th, 2011 at 12:13 am
govt_mule
“You cut corners in every way possible, you do anything to push the price down. Send it out half baked, full of holes and ugly as hell, but just send it out cheap.”
Microsoft in a nutshell. Why didn’t that model work for Yugo? Alternatively, do shitty, undependable products really save users money in the long run?
Tuesday ~ October 18th, 2011 at 6:03 pm
Gene Callahan
Curt, what does whether a language has strong or weak typing have to do with whether it is interpreted or compiled? You can have strong typing in an interpreted language, and weak typing in a compiled one.