Friday, August 6, 2010

A fund manager experiments in geek...

Forgive a long post – but I suspect this might develop into a theme.  My business partner thought I should break this post up – but I figured I should put the outline in one place and refer back to it.  There may be follow ups but this stands on its own.

My original motivation for this examination was Whitney Tilson's rather convincing looking buy case for Microsoft.  Microsoft is – by all conventional measures – pretty darn cheap these days.  Whitney Tilson (a well known hedge fund manager) details how cheap and he does so without mentioning what I think is the most important driver.     

The driver that I think Whitney was not mentioning – and which I have not quantified – is the shift of much of the world to laptops.  In America or Australia if I buy a desktop PC it almost invariably comes pre-loaded with Microsoft.  It is however a surprisingly easy thing to do to build your own PC – buying the components separately to get precisely the PC you want.  I have done it and works.  The reason this is possible is that standards have made the PC modular – all the plugs and protocols are common.  Nobody other than serious geeks (or gamers wanting to soup things up) does that in West – but in developing countries it is not uncommon for your local computer supplier to do it.  After all – if you put a computer together yourself you do not pay anything to Microsoft – and you can – if you want – you can install a pirated computer system.  By contrast. if you go to a computer shop in Mumbai there is a fair chance that the machines are “white label” and the system is “not genuine”.  However there is almost nowhere in the world you can buy a (retail) laptop (other than an Apple laptop) which does not come pre-loaded with Microsoft.  You used to be able to buy “bare bones” laptops in Australia – where you added your own memory and CPU to your own specs – but I can't find them anywhere.

As the world (particularly the newer markets in India and China) moves to laptops the “not genuine” problem for Microsoft evaporates.   Growth in laptops in India is nirvana for Microsoft. 

And Microsoft is so cheap that any increase in sales will make the stock look extraordinary.  The stock at this point does not need many drivers.   

The negatives are the re-emergence of Apple as the “must have” product – especially for the young and affluent and the encroaching of linux – which is now dominant in servers – as well as entering the retail market through things like the $250 net-book computer and through being the basis for some mass market products (ranging from Android phones to PVRs).

I wanted to explore the negatives (especially linux and virtualization) properly and so I spent a couple of weeks turning myself into a geek.  This post explores that journey. 

If you are a “real geek” and you know what you are doing you are probably going to find all of this shockingly naïve – but I still want your comments because that is how I learn.  If however you are like me and pondering the stock maybe I will save you some work. 

I will start with a conclusion which should not surprise any geek – but tends to surprise non-geeks:  linux is the “real deal” and is a much bigger threat to Microsoft than Apple.  However it will also change Apple's (laptop) business model beyond all recognition – and it will do so via virtualization.   It will also change the hardware business beyond recognition.  Indeed it is already doing so.

I have now changed my laptop to a linux (Ubuntu) machine and run a piece of software (Virtual Box) on it.  Virtual Box is a program which pretends it is another computer – a virtual computer.  On virtual box I run Windows.  This is – I believe – a superior set-up and it is unlikely I will ever run a machine primarily on Microsoft again.  I will explain why more fully below – but first I just wish to make a simple observation...  if I take the hard drive out of my laptop and install it in my old laptop everything works just fine – the whole computer is functional.  If I tried to do that with a windows operating system it would fail.  This is likely to be important in the future of computing because I will be able to migrate my computer from a laptop to the cloud – or possibly onto my (linux powered) phone.  It is unbelievably useful to have a hardware-independent computer. 

I warn in advance that I have not come to very strong stock-conclusions.  But I have learnt a lot – so this is a post to detail what I have thought along the way – and an invite to have the “true geeks” correct me. 

Three business models for an operating system

There are three business models around for operating systems – two of which are closed source (Apple, Microsoft) and one of which is open source (Linux, FreeBSD etc).  The first part of this post explains these models and hence explains what software is good for and what it is not and how that derives from the business model.

Historically the most important of these models is Microsoft.  Microsoft builds software which they license very broadly.  Anybody is allowed to build a computer and run Microsoft on it.  Microsoft sells to almost all computer manufacturers.  Hardware makers competed to produce better and faster and cheaper computers.  All of that competition took computers from a highly niche product in 1984 to a must-have for a very large part of the world.  Microsoft split the benefit of all that competition with their customers and became frighteningly profitable – probably the best business in the history of capitalism. 

But Microsoft's willingness to work with any hardware makers is also the big weakness in the system.   I have a computer assembled from bits found on Ebay and (sometimes literally) from bits at the side of the road – and it runs Microsoft Vista well enough.  It is probably different from any other computer in the world – but Microsoft has to make it run.  And that is difficult for Microsoft – and the computers are by-their-nature buggy.  This is meant to work with that – but conflicts are rife and sometimes computers have errors for unidentifiable reasons.  The reasons are unidentifiable because Microsoft does not release their source code and without that nobody (and I mean literally nobody) can actually work out what went wrong in some instances.  Microsoft can – but only because an “error report” gets sent back to them and – provided they get around to your problem – which presumably means provided the problem is widespread – they can issue a patch. 

And that roughly explains Microsoft's position in the market...  it works for everyone – but it does not work particularly well (though Windows 7 is an improvement).  Moreover Microsoft allows you to use cheap hardware – and that is a good thing. 

Apple deliberately took a different tack and the company almost failed.  However that tack has resulted in Apple’s new resurgence.  Apple do not license their software – and so – because they built every computer allowed to use Apple software they know precisely “what is in the box”.  Because they know “what is in the box” the machines work.  After all – they are not building for 100 thousand different configurations – possibly only the thirty or so configurations that they have sold.  And they can test the software on each of those configurations and if it works there they know it works.  (They also limit multi-tasking on some devices like the iPhone because multitasking with 100 thousand apps will produce combinations that they cannot possibly have tested.)  Less configurations means less complexity.  And because of that there is less “bloat” in the system which makes it faster (for any given hardware). 

But there is a downside with the Apple model – and the downside is that there is less competition between hardware markers.  Competition between hardware makers works for Microsoft far better than it works for Apple and it meant that Macs were always over-priced (even allowing for the fat margins that Apple builds into its hardware).  To some extent Apple solved this by moving to x386 (ie Wintel standard) chips – and allowing the Microsoft competition to work for them.  This made the machines cheaper but also allowed geeks to load Apple software on non-Apple machines (ie making a “hackintosh”).  

This model positioned the Macintosh in the market.  It was the computer that “worked better” and was not glitchy – but it was a niche product because it was more expensive. 

It was on this comparison that Microsoft rolled over Apple and became the dominant and most powerful computer company in the world.  Macs – it seemed – were doomed to be the (darn nice) niche product. 

When Apple finally moved to x386 chips the difference in production costs became less stark – but they remained – and they remain to this day.  Still – the fact that Apples work means that for many uses the total-cost-of-operation for a business running on Macs is lower than a Wintel setup.  I know a medical centre that recently changed and considerably cut costs.  Further – and this bears observation – computer hardware is now getting sufficiently cheap that the disadvantage of the Apple model (lack of competition in hardware) is becoming less significant. 

It is worth understanding how larger businesses (say 200 plus computers) have dealt with the problems of Microsoft.  Essentially they have tried to give Wintel platforms the advantages of Apple platforms via standardization... 

What they do is rather than have 200 different computers throughout their business they have maybe one, at most three different models in use at any time.  These computers are essentially identical and they have – sitting on the IT guys desk – three exact clones.  When new software arrives (say for example a Microsoft update) the software is thoroughly tested on the three clone computers to make sure it produces no glitches.  If the software (or hardware upgrade for that matter) causes no problems they roll it out across the network with all the computers being changed when staff power them up in the morning.    The system works because the IT specialist controls “what is in the box”.  By controlling what is in the box (often restricting the right of staff to load their own software) they get Apple levels of reliability but the ability to buy Wintel priced hardware.  They do however pay a price on hardware – which is that they often get tied to exact specifications for computers.  If their business expands they can either (a) get a new computer specification in or (b) order more computers under the old specification.  When they do the latter (which would be most the time) the hardware maker has leverage – and selling old computers to business at old prices can be surprisingly profitable.  [After computers should fall in cost by about a percent per week – so selling an old computer at an old price is a massive winner for the vendor... and is an important part of Hewlett Packard’s business.]  Big business – through standardization – are – in this view – trying to emulate the advantage of being Apple.

There is a third model out there – which is the truly open-source model.  In open-source the source code for the software is public and so – with appropriate expertise anyone can see what is going on in the software.  [This is different from both Microsoft and Apple as their proprietary software makes computers running on them into true “black boxes”.]  If you know what is going on in the software you should (again with appropriate expertise) be able to make your hardware work with it.  If your hardware causes glitches you can fix it (or fix the drivers) because you can see what is going wrong.  The hardware makers have both the means and the incentive to make sure it works.  With Microsoft and Apple they have the incentive but not the means (they can't see the source code).  So Linux is strangely the best of both worlds – there is competition between hardware makers so it uses cheap hardware and uses it well and it is really stable – as stable as Apple. 

There is however a downside – which is that frankly – nobody much has an incentive to make the whole thing work for the end retail user – that is – the user interface has generally sucked.  The thing now comes with two user interfaces (Gnome and KDE) and nobody has much standardized anything.  Moreover the interfaces are run by committees and – by the standards of open-source products – they are bloated.  This downside is rapidly been remedied.  A couple of years ago I tried using linux on the desktop (using Mandriva and OpenSuse) and frankly it was not much chop.  I now use Ubuntu and it is nice (meaning moderately friendly).  The remedy however is slow at coming.

This description places Linux in the market too.  The description is (a) stable, (b) able to use any hardware and (c) less-attractive and user-friendly around the interface.  This makes it perfect for nerds and geeks who like the stability and like being able to run their peculiar hardware setups and don't care that it requires some expertise.  The place where linux found its main home was on servers.  Servers are computers that run all the time set up by geeks and let run without much attention.  They tend to need to work together.  They are perfect for linux.  Microsoft has slowly and surely been losing the server market to linux.  If you need your server to work really reliably with hardware of various ages it is almost certainly running some form of linux (or maybe FreeBSD another open-source alternative).   The Googleplex unsurprisingly runs on linux derivatives as does the space station and probably most nuclear reactors.  Its a great system but user-friendliness has always been an issue.  [I know that the linux geeks will object to that statement – but it is odd having to learn codes like “sudo apt-get install cinlerra” and discovering it does not work…]

One more thing that linux offers is considerable ambivalence about which hardware you use.  You can run it on many kinds of computers.  If you take your hard drive out of one computer and install it in another computer it will probably work straight away.  Both Apple and Microsoft are very hard-ware specific.  This willingness to be used on many computers is a massive help for a system administrator because she can update the hardware and it will work.  Compare this to when you buy a new laptop and you need to reload everything on your shiny new machine.  Hardware portability is anathema for Microsoft because Microsoft sell new software every time someone buys a new machine.  I replaced my laptop because well – my Dell machine was a turd – and Microsoft – bless them – extracted over a hundred dollars from me.  I was not replacing the system which was adequate – I was replacing the machine which was inadequate.  Hardware independence would have been incredibly valuable to me because there would be much less problem with migrating settings and other painful but essential tasks. 

Apple is more comfortable with hardware independence because you are always using Apple hardware into which a massive margin is built.  Microsoft make their money selling software – Apple by selling hardware at fat margins.  You know this because you can buy an Apple operating system for $40 down at the local shop – and it is a better system than Windows.  The only problem is (at least according the end user license*) you are obliged to run this on an Apple machine.  Microsoft sell licenses for their system for about 8 times this sum. 

The ipod, iphone, Ubuntu, virtualization and other business model changes

There have been several large challenges to the outline I described above.  These were not all that predictable – or at least if you predicted them you might have made a fortune on Apple stock.  I will deal with them in order.

Firstly Apple have found a niche where their product is simply superior.  It is consumer products for non-geeks that want to be (a) super reliable and (b) easy to use.  The first of these was the iPod.  It was a dead-easy to use music player that met an enormous consumer desire.  The previous products in the space (portable CD players, Walkmans) were – at least by the standard of an iPod – very inferior. 

The thing about a music player or a phone is that it is ultimately not-that-expensive and it is really important that it just works.  And so the Apple model is just superior.  And you know this – how would you feel about having Windows on a phone?  The question answers itself – there is no reason for a buggy thing (or even a thing with a reputation for bugginess) on a product designed for limited functionality and unsuitable for hardware expansion. 

The other non-buggy operating system (linux) is also suitable for phones and Google has done it – the so called Android – which is really a dressed-up version of very-small-linux. 

The niche here for Apple however is for products that are resistant to hardware expansion and have to just work.  iPads, iPhones and others fit.  The completely closed software shop is part of the way to get these things to work.  If the software shop is completely closed then you know it won't cause glitches because you control (a) the software and (b) what is in the box.

Windows 7 might be modified for a tablet PC but it is likely you will think about Windows on a tablet just like you think about Windows on a phone?  Why bother? A linux limited purpose computer makes more sense – and that is what most netbooks are.    The $200 netbook comes with an operating system and some basic software (web browser, word processor, spreadsheet etc) – and with little expectation by the customer that they will dramatically expand their use.  A super-stable open-source system is just fine.  [You do however attract consumers by allowing games and fun-stuff – and the Apple software shops do that better than an open-source platform.]

The second big change is that the front-end of linux is now becoming more user-friendly.  Ubuntu is the big driver here.  Ubuntu is a distribution of linux originating in southern Africa with an explicit aim of user-friendliness.  Also some of the key products (for instance Open Office) have crossed the threshhold where they are as nice and as functional as the Microsoft equivalent.  For most people Ubuntu is a superior operating system to Windows.  It is less bloated, does all the key functions and is more stable.  Ubuntu has made netbooks at $250 possible – it is a fully functional operating system that will work on a cut-down computer and can be distributed without paying the pound-of-flesh to Microsoft. 

There are however problems.  The first is several bits of key software will not run on Ubuntu/Linux.  For most consumer uses the show-stopper is iTunes.  Apple produces a version for (inferior) Microsoft but will not port it to linux.  My guess is that doing so would seriously undercut their own business because – frankly – linux has the stability advantage of Apple at a fraction of the price.  From my perspective the difficult programs are Windows Media Player, Windows live writer and my feed reader.  There are several open-source products which will play Windows streams.  However when the streams become heavily featured (with interactive slides for instance) the open-source players simply fail.  I have learnt to hate companies that do their conference calls using proprietary Microsoft systems.  [The biggest offender is Bank of America.  Will IR please stop it.  Pretty please you slime-bag monopolist lovers...] 

Still I can't expect Bank of America IR to stop using Microsoft proprietary systems for their conference calls until there is a critical mass of people who complain or cannot listen – and there will not be a critical mass of people who complain until there are enough people who go open source.  We have a critical mass problem.  You are forced to continue to use the inferior Windows product because – well – everyone uses it.  And everyone else uses it for much the same reason.  The $250 net-book is a serious threat to Microsoft because it might over time produce the critical mass of people who use Ubuntu or like systems – and that will enable us all (even those of us with pricey laptops) to switch to Ubuntu.  [It is a threat that Microsoft is meeting by allowing cheaper operating systems on super-cheap laptops.  You can buy a Window 7 laptop in Aldi in Sydney for the same price as a software license.  Obviously Microsoft has discounted somewhere…]

There are also other costs of using Ubuntu.  The main one simply being that things are different to what you are used to – and hence difficult.  For instance you go to the “start” button in XP to turn it off.  Who would have guessed that?  But you “know” it instinctively.  Likewise you know that there is a “snapshot tool” in Vista – but you do not know that the equivalent in Ubuntu/Gnome is “shutter” and you need to download it as it is not standard with the system.  We have many years of human capital invested in Windows – and even though it is “inferior” it takes some weeks to change.  It is not a light consideration.  Inertia is a powerful thing.  (Mind you Microsoft faces its own inertia as it tries to move happy-enough XP customers to superior Windows 7.  Inertia cuts both ways.) 

I strongly suggest you (dear readers) solve the inertia problem for your family and give your 7-10 year old kids old computers loaded with Ubuntu.  Relative to anything else you can give them on an old machine it is mondo-powerful – and they will grow up understanding computers in a way that you simply do not if you are not a geek.  Open-source is a superior learning environment.  [My son loves Ubuntu...]  Any school teacher who runs a primary school computer classroom where the computers are not Ubuntu is – frankly – being lazy.

There is a reason why Ubuntu has suddenly got better though – and that there is a rich individual – and more recently Google is behind it.  There is an internal Google operating system (not publicly released) called Goobuntu.  Security concerns mean that Google staff are now prohibited from using Windows.  More pertinently the new (highly minimalist) Google operating system (the Chrome system) is a cut-down version of Ubuntu.  Google intends on using Ubuntu and its derivatives to hammer Microsoft – and – frankly – the faster your children learn to use them the better.  [The cost for an individual changing is high – I reckon about 4 weeks of productivity – about what it has cost me... but the cost for a child is zero because they have no human capital built into the alternative system.]

There is a third major change to this operating system business – and that is “virtualization”.  Virtualization is the business of running a machine within a machine – or for that matter a machine across several machines pretending it is one machine.  For instance when you query Google it seems like you are querying one machine – but in reality the Googleplex is maybe a million machines pretending it is one machine.  Similarly one machine can be running six to 100 virtual boxes on it.  This will fit many of my readers.  If you ran a financial business with say 100 staff with their own machine the way you would set it up is with four (powerful) servers – two in the main office – and two mirrored machines in the remote back-up location.  Each staff member would have their own “virtual machine” sitting on the paired servers.  They would have allocated RAM, processing capacity and hard drive space but when that is not been used it would be allocated to other staff members.  Every virtual machine could be made available off-site (for example if staff members travelled).  When staff change their desk their machine (which is virtual) does not need to be moved.  More to the point – the machine is entirely hardware independent.  If you need more RAM collectively you just add it.  The servers would be running some flavor of Linux (probably SUSE or Red Hat), the virtual box would be either open-source (“Virtual Box”) or proprietary (VM Ware) and sitting on the virtual box would be Windows or Ubuntu or – for that matter – Macs.  [I will discuss virtual Macs later on...]  The computer can be migrated from one server to another dead easily [the “hardware” is the virtualization program].  It can be scaled easily.  It can be duplicated easily.  Moreover the system can be made generally redundant easily (the Googleplex has much built-in redundancy – if a computer or a thousand computers in the Googleplex goes offline it does not much affect the service).  VMWare is arguably the hottest stock in the hottest sector at the moment. 

The beauty of hardware independence is that everything can be changed – and by running (extremely stable) linux and (unchanging) virtualization programs you can bring the stability of linux to everything.  The virtualization set-up is frankly superior – and it will improve the stability of Microsoft – perhaps eventually to Apple levels. 

And that sounds fantastic for Microsoft – but alas it comes with a very big price ticket.  Microsoft relies on hardware dependency for sales.  The reason I buy a new operating system is that my Dell sucked – not because I really wanted to own Vista (or even Windows 7).  I was happy-enough with XP.  I had to upgrade simply because – well I had to upgrade my hardware.  The motive for buying a new computer is almost never because it runs the latest version of Windows.  The motive is that it is a bigger, more powerful computer and my old one can't keep up with my demands (or is defunct as per most Dells).  Indeed there is a cost to upgrading.  [I hate the new picture-driven menus on Office 2007 – and have reinstalled my old Office 2002 because I am used to it.  The upgrade sapped productivity and gave me the incentive to learn Open Office.  After all – if I have to learn it all again I should – by rights learn it all again on something that is free and possibly superior...]

Virtualization – and hence hardware independence – will simply mean that Microsoft sells much less.  Indeed – I can't see why they need to sell anything at all after they have sold you a virtual seat – you can just upgrade the hardware around your machine and keep the software as pristine as you like.  Indeed if the server you use is out there in “the cloud” you will never need anything other than a terminal.  Virtualization – and hardware independence – is really scary for the boys from Redmond.  [By far the most unconvincing argument in Whitney Tilson’s piece is that Microsoft is a key player in the new trend of virtualization.]

And Microsoft know it too.  A while back Microsoft was telling us it wanted to change its model for selling the product to business.  Previously they sold it on a one-off license basis.  Now they wanted to rent it.  And well might they – because with virtualization you might go a very long time between upgrades.  They know that consumers (who buy machines rather than seats) would not be interested in that – but maybe they could sucker business along with a lower first-time charge. 

Apple and virtualization

This is a harder topic.  Apple by-and-large do not sell to “the enterprise” and their computers do not virtualize that easily (primarily because in most countries they will not let you).  The Apple End User License Agreement (EULA) makes you agree not to install the software on any non-Apple machine (except in countries where such a restriction is illegal).  I think Australia is the most notable “except in” country – which means (I think) I am allowed to install the software on a non-Apple machine.  [Third line forcing rules in our antitrust legislation would make it a criminal offence for Apple to enforce their license terms...]  Anyway I set up Snow Leopard on Virtual Box – and to tell you the truth – installing it the first time was a first-order pain.  However copying it would be easy.  [After all – it is virtual and I can copy it very quickly – it is just hardware independent software...]  I am going to close it and give my snow-leopard disc as an upgrade to a friend because – frankly – now I run Ubuntu a Mac is simply not that attractive. 

Apple don't sell much to “the enterprise” and their model is to sell the software cheap (the system cost me about a tenth of a Microsoft system) and make money on retail sales of hardware.  If they can get Apple into big business they win – and they are not savaging their own hardware sales.  Apple might get to sell some of their (outrageously expensive) server products to companies that might virtualize.  And the given you only need to buy the seat once – and the young customers love their Macs (for good reason) it might actually be the sensible way to run.  They might also sell virtual macs through the cloud – at a rental fee. 

I suspect however there is another game here.  Hardware independence is a truly wonderful thing for mirroring.  When I take my Windows hard drive out of my Lenovo and put in the (crappy) Dell it does not work.  But my linux disk does.  I can set up a mirror for my laptop to a desktop at work (as the same information will work in both places) – so I do not need to cart the laptop with me to and from the office.  Given how fast computer processing is getting small and powerful there is a reasonable chance I should be able to clone a whole computer into a mobile phone – and keep it in my pocket – but have it securely running on the desktop as well.  Apple could win at this – and I do not want to speculate as to where they are going.

So back to Tilson’s argument

Microsoft clearly is really cheap.  And there is a huge driver he has not even talked about – the shift of (say) India from white-label desktops to “genuine Microsoft” laptops.  But I suspect the whole business is more vulnerable to a complete paradigm shift (virtualization, cloud etc) than I would like.  There is a chance Microsoft’s business just collapses – and with it the hardware business.  Hardware independence really is the big deal and whilst I might consider the stock I would not label the holding permanent.  The boys from Redmond should be scared because they rode the wave of distributed desktops and laptops to glory and that wave looks a little stale now.




PS.  In doing this someone suggested to me that if you play a Microsoft Midori (their future virtual offering) disk backwards it asks you to pay homage to the Great Satan.  He however suggested it was far more dangerous to play it forward.  In that case it might actually install Microsoft Midori. 

PPS.  Whitney… if you have made it to the end of this then I am saying hello.


Zardoz said...

That's an interesting and useful survey of the current situation. And that link to VirtualBox is priceless. But I can't help feeling I've missed something basic in the argument.
Remember that capitalism is about combined and uneven development. Why won't all the Indians using laptops with fake Microsoft OSes just jump straight to netbooks with real Ubuntu? I can see some of the argument about sunk human capital costs in learning the OSes but device cost must be a big factor in poorer countries?

John Hempton said...

MSFT strongly discounts to get their stuff into netbooks -

Here is a 389 dollar netbook running windows 7 in Australia. The price of a raw system - just an install CD and code - is $299.

Pretty clear that the Softie margin has been cut sharply.

I think MSFT will discount in India to solve the same problem... but they will still collect some revenue where they currently collect none...

matt said...

As looking at this from one remove - a fund manager (though structured FI) with repressed geek tendencies (large numbers of small children having taken up the available time) - I'd add:

1) More of the computing space is up for grabs to phones and other quasi-computers (e.g. the IPad and IPod touch) than make me comfortable taking a large position among the traditional PC makers (including that side of Apple). It's hard not to see the majority of non-work (and non-schoolwork) computing moving to devices of this sort. They are too inexpensive, portable, and easy to use to be avoided in comparison to laptops. Microsoft does not look good here.

2) Desktops have rebounded for work purposes thanks primarily to low cost multiple monitor setups. The productivity gain of having 2-3 cheap LCD monitors over a small laptop screen is large. Running the office systems (virtualized over Citrix) and Bloomberg Anywhere (likewise) one window at a time on a laptop screen is certainly not pleasant or productive. Microsoft does look good here; neither Apple nor (as far as I know) Linux have the range of enterprise tools to be a strong choice for medium-large non-tech businesses. But there is not strong growth here.

Shooter said...


Intersting topic, myself i would be selling MSFT, if not shorting them (for a very long term view). They are the Telstra of the Computer market. Old school management with old school ideas.

Open source is going to be the future, just look at how it has grown and it holds to true values of the internet close to heart. It's free, highly compatable, adjustable, and no one can really profit from it.

Google can see this and just look at everyone of there products. They follow suit, only they have the business model to maximise from this and try to get the market to access the world through there products, whilst advertising on the side.

India wont go for cheaper MSFT offerings, they (as well as other aisan countries) are so advanced in this area that they will be on the Open Source products, if not already. An example is how quickly the Asian population addapated to social networks like twitter, how accessable free wifi is (or how cheap there phone contracts are).

A bomb goes off in an Indonesian hotel and the picks are on twitter within seconds! And its all for free.

Buy the business with the stratedgy to make money on the future direction of the indutry. Sell the company that made it all possible but did not stay ahead of the game because of over protecting there product.



HSNYC said...

I have spent a long time looking at this company over many years. I can confirm that there is plenty of discounting with Windows (subject to the anti-trust agreement limits). OEMs receive volume-based discounts (which gives them advantages vs. start-ups and incentives to avoid Linux). In addition Microsoft created a stripped down version of Windows to sell in emerging markets (read China & India) to compete with themselves (that is pirated Windows). This stripped down version is a fraction of what Western version costs. That said even the fuller OEM version is no more than $50 for the major OEMs.

I have heard many insightful and intelligent comments about the industry and where it is headed. After a while you realize that no one really knows. But just for fun here are a few insights some of which you have already stated:

• Bill Gates was masterful at economies-of-scale, customer & supplier captivity and tie-ins.

• They came of age when it made sense (from an economies of scale perspective) to have one company make sure that everyone's components worked well together.

• The way you lowered the cost of the PC was by having a lot of competition among the component makers. Also suppliers only want to write a driver for one OS (supplier captivity).

• This was because we all wanted better, faster computers to do tasks at the cheapest price (OS cost was small relative to the total cost of the PC).

• As a result Microsoft's culture was built around replication (and tie-in) of what other companies had already done and not innovation.

• Due to their success and competition among the suppliers, Windows & Office are now a huge part of the cost of a PC (mostly because the component pricing has fallen).

• Since the total cost of a PC has fallen so much, Apple has been able to make inroads by selling a total consumer product (advantage is obviously less bugs at a markup).

• Microsoft’s big headwind is the declining mind share and utility of their office products.

• Right now people continue to use Microsoft products because they are used to them and because of the formatting issues around using anything else (customer captivity). Additionally, businesses continue to use Windows because they have already built applications and interfaces that work on Windows and it is a huge pain to have to rewrite all of the code.

And finally I will add that I have heard that when Buffett was asked by an employee of Microsoft what he thought of the business, he said the most difficult thing is deciding how much to charge for the product.

IF said...

Ugh. I don't know where to start. What you write in general is reasonable, but slightly off. Too much to address here. For the high level picture about MSFT I agree. But for different reasons.

It seems MSFT is pretty good about enforcing "Genuine" Windows these days. This is pushing the culture in many threshold countries towards Ubuntu. Most people in Bulgaria seem to be using it. We gave a gift of a new Win7 netbook to the 65 year old parents in law and they made sure it got blown away and Ubuntu installed.

I used to run Linux for many, many years. All my undergrad and PhD time as a computer scientist. Have not touched Linux in 5 years. XP was okish. Vista a disaster. Win7 is fine. And stable. I don't know where you get that "unstable" from.

There are different versions of Win7. The one for 299 bucks is probably Ultimate. The one on the netbook is Home Basic or at best Home Premium. MSFT will charge later for an online key to upgrade.

You can move a Windows drive to another machine. What is preventing you from doing so are the hardware drivers installed. If you remove all of them in safe mode before changing drives your chances of success are much higher. Worst case use an install DVD to "repair". Linux probes all drivers dynamically during boot. If you bake them in you run into the same problems.

[to be continued, comment too long for Google]

IF said...

[Part 2]

Windows 7 Ultimate has virtualization built in. Also, notice that many software companies used to charge at least since the 1960s depending on how fast your computer is. What should prevent MSFT doing the same with virtualization?

Some people already argue that laptops are dead. Long live androids and iPhones... A problem for my company.

Apple does not use the limited hardware choice to make things fast. At least for 3D performance you will usually do better installing Windows. Their drivers and middle-ware try to ensure that everything really, really works. Let me repeat, Apple does not care for performance.

Yes it is a problem to have closed source, even when working closely with MSFT and getting OS drops every two weeks during development. We are working in their kernel mode and life would have been so much better with sources. On the other hand Lunix has the opposite problem with some modules/drivers (that will not publish their sources).

Personally I wish Linux the best. Don't care about Apple. I recommended a friend to buy Asus Win7 laptops that are faster than Apple at half the price. (At which I consider them pretty much being disposable.) His Apple came physically warped and he tried to get a replacement. No luck in the whole Bay Area, as he ordered slight custom screen resolution. They are currently shipping one from Shanghai (two weeks of waiting so far for a 2k laptop that was meant to make him "feel good" over a cheap and ugly Asus).

But I do think that MSFT might be unstable. Maybe a 1/4 chance of disruption in the next 10 years?

Anonymous said...

your take on emergent China and India laptop demand is interesting.

Of course the street cnsensus is that MSFT's P/E would re-rate upwards once corporatations start their Win 7 upgrade cycle.

But waiting for corporate Win 7 upgrade is akin to waiting for Godot. It hasn't materialize yet and I'm defintely not feeling the love.

Apple is defintely the Sony of the new mellenial. Their core business is now selling consumer appliances (music players, tablets and phones).

Phil said...

Bear in mind that (as I understand things) the big profit driver for MSFT is Office and services (a la IBM), The OS is to some extent a loss leader for the Office hegemony. As long as businesses feel that they depend on the availability of Office, then they'll keep on paying the MSFT tax on their employees.

Whilst there are perfectly functional Office competitors, none of them can guarantee full compatibility & so Office has huge inertia. The only thing that will shift it would be some kind of mass shift in end-user usage. A move to cloud-style productivity apps perhaps? MSFT has already recognised this potential threat & is developing ways to deliver Office in this fashion though, so I wouldn't write it off yet.

Even if you effectively regard MSFT as being in "run-off", it's going to make a huge amount of money from Office sales over the next decade or so. After that, who knows?

狂猪 said...

Part I ( comment limit)

1. There is another very important advantage the Microsoft model has over Apple (a potential short opportunity which I discussed in #5 below). Because Windows worked with any PC manufacturers' machine, the marketing dollar of every PC manufacturer is in affect also marketing Windows. Notice all the Windows stickers on every PC? PC manufacturers time their production plan with Windows releases and marketing campaigns. Whatever marketing dollar Apple puts behind the Mac is a pittance in comparison to the aggregate marketing power of all the PC manufacturers and Microsoft. My description of this relationship is really not doing justice to how powerfully it has benefited Microsoft over the years.

2. Running Linux is too demanding for an "average" user; never mind running Linux+Virtualization+Windows. I've been writing server software for UNIX variants for 15+ years and I run Windows at home. Linux still lacks in hardware compatibility and user interface design. My experience has been if you have hardware peripherals that is really new or if it is not a popular brand, you'll likely have difficulties with Linux. And with Linux, anything slightly complex quickly degenerate into editing some esoteric config file. Linux is great for servers and lousy for the average user.

3. The OS isn't what keeps an average user on MS Windows. It is mainly MS Office. More generally, it is the size of the software library for Windows. Microsoft will never port MS Office to Linux. For most other software vendors, this is a chicken and egg problem. Most software vendors cannot afford to be the first mover. It is possible to run Linux+Virtualization+Windows. However, realistically, for the next few years, this small user base is too small to over come the chicken and egg problem.

狂猪 said...

Part 2

4. The Apple model isn't an advantage for mobile devices. PC hardware compatibility generally did not came from PC manufacturers. Manufacturers test and make sure all the bits inside a box work together and with Windows. The compatibility problem is from add-ons and from people who build their own computers from parts. For example, people often swap the graphics card for better ones to play games. Hardware compatibility is not an issue for mobile devices. Every phone leaving a factory will have been fully tested and there aren't that many hardware add-ons. Upgrading the OS on an old phone is a different story. Here, even Apple cannot upgrade the first gen IPhone to the latest IPhone OS.

5. Apple today is repeating it's Mac strategy from the 80s for it's IPhone platform against Google's Android. This decidedly looks like a movie we've seen before. In time, Android devices will cost half as much as IPhone, run 2x faster than IPhone, and the Android platform reaps the benefit of the aggregate marketing muscle from all the Android manufacturers and Google. Today, the Apple market cap is $240bn. Apple had an extraordinary run in the last 15 years. Mean reversion dictates companies don't grow to the moon. Furthermore, fashionable things tend to stop being fashionable when everyone has a copy. Even Walmart is now selling Iphones (reminds me of Motorola Razr)! Doesn't it seem the technology landscape, probability and statistics, and even fashion trend are screaming "Short!!"? John, what do you think?

6. MSFT is a $220bn market cap company. Statistically, companies as big as MSFT have very difficult time outperforming the sector they are in. I will not long MSFT. I rather pick winners from the little guys and pick loosers from the big guys.

Disclosure: no position in MSFT, GOOG, AAPL

cturner said...

You call yourself a non-geek but that was a sophisticated analysis. Nice work.

Unix is an addition. The more time I spend with it, the less time I spend out of console windows. The console is tremendously liberating.

Andrew London said...

Probably should point out that Apples software is a fork of FreeBSD:

not a bad analysis at all.
Look at these 300 million dollar "social gaming" companies. Platforms are where the money is, not software, thats what the whole "cloud" (i.e. the internet) bullshot is about in IT circles at the moment. it's what the internet / IT has always been about, repackaged as a chargeable service.

Anonymous said...

I run an Ubuntu PC and a MacBook and both are stable, user-friendly, and trouble-free.

Some Windows apps which I can't live without, I run under WINE. (I have iTunes on the Mac and don't need it on the other machines, but some of its functions work under WINE:

I use Firefox, Thunderbird, and Open Office cross-platform.

The only reason I need Windows is for Bloomberg - and to link spreadsheets to Bloomberg I apparently need Excel. (If anyone can link Bloomberg data to OpenOffice spreadsheets, I would love to hear about it.)

So, I do have one Windows PC in the office (running Windows 7, which is proving a real pain). I also run Bloomberg on WindowsXP on the MacBook, using Parallels Desktop, which works like a dream.

When I can get rid of Microsoft products altogether, I will be very happy. Ubuntu is a delight. I would second your recommendation of it for everyone outside the corporate Windows environment.

Felix said...

A pretty good summary for non-geeks indeed.

I agree with you that Whitney underestimates the impact on MSFT from threat from open source and free software that is supported not by license sales but by services.

His charts all show *revenue* share. His charts only show that competitors have not been successful at stealing revenue share from Microsoft. What they have successfully done is start to shrink the pie.

Licenses are much higher margin than services. Microsoft can keep 98% of the pie all they want, but if the pie starts shrinking materially, earnings will take a material hit.

This is unique in the software industry. A competitor can kill the market without ever making a single dime of revenue. All they have to do is crush pricing. Ironically, this is how MSFT killed a lot of their competitors. For example, MSFT killed Novell by giving away network software. MSFT never made a single incremental dollar from its network software. Novell's business model evaporated overnight.

michael webster said...

Very interesting analysis. But, I think that the review of the 80's is flawed. There are several important omissions to the OS story.

1. Why did OS/2 not catch on? Why did all the niche OS's die? Amiga, Osborne, etc.
2. Why was Microsoft able steal software market share from the big word processors and spreadsheets?
3. How did Microsoft get away with so many vaporware announcements to kill the competition.

But I did like the post.

F. Heinsen said...


Hi – I've been subscribed to your blog ever since I came across your post about overpriced hookers in Riga. I'm commenting for the first time, as this post falls in my circle of competence.

IMO several important issues were overlooked in your post:

First, your comment implying that Linux user interfaces are bloated because they are “run by committees” overlooks the fact that virtually all free, open-source projects are managed in roughly the same way, and that these “committees” are really complex online communities that make software using a different approach than the one used by proprietary vendors.

If you follow any of these communities for even only a few days, you will find constant dispute, complaining, scheming, accusations, dire warnings, etc. as they try to work things out in the open. It's not unusual to see adversarial factions within the same community trying to solve the same problem with competing approaches. Sometimes multiple solutions are officially adopted. It's sink-or-swim, really.

To someone unfamiliar with the process, it looks like utter chaos – inefficient and slow.

But if you think about it a bit, you'll realize that what we're looking at is a survival-of-the-fittest, evolutionary approach to software development. Free, open-source software is the product of evolution, not intelligent design. That is why and how Linux runs so well on virtually all hardware -- the same code on everything from phones to netbooks to PCs to supercomputers.

Applications like Firefox and OpenOffice are all gradually evolving to run everywhere too. And the underlying code is not subject to obsolescence – code published under a free, open-source license can never be withdrawn from the marketplace. So long as someone needs the code, it will be used – and evolve.

I'm not sure Microsoft's 'intelligent-design' can successfully compete against evolution.

Second, as hardware costs continue their inexorable decline, the lowest-cost software platforms -- i.e., Linux and its relatives -- become more compelling. India recently announced a $35 iPad clone ( This price point is possible only because the device runs Linux and other free, open-source software. How long before we see a $10 (actual, non-subsidized, cost) tablet?

I'm not sure Microsoft's high profit margins are sustainable in the long run.

Finally, and this is a corollary of the first two points, the most important long-term trend I see in the software industry is that the most-compatible, lowest-cost platforms -- i.e., Linux and relatives -- are gradually expanding into every little corner of our computing infrastructure.

My Motorola cable-TV boxes run Linux. My Tivo runs Linux. My Droid phone runs Linux. My Control4 home-control system runs Linux. My Sonos audio system runs Linux. My Cisco Internet router runs Linux. My LinkSys WiFi hotspots run Linux. My network drive runs Linux. Recently, on an international flight, I noticed that the in-flight entertainment system on every seat was running Linux. All the online services I use run on Linux or its relatives. And yes, my desktop PC and my laptop run Linux too (Ubuntu of course).

Market share statistics, which are based on sales figures, DO NOT reflect the widespread adoption of Linux because free, open-source software is available at no cost. Consider only that Google, Amazon, FaceBook, Twitter, Yahoo, and many other giant and less-giant websites have many millions of server instances running -- for which they paid exactly $0.00 in OS licensing fees.

In short, almost no one has noticed, but to me it's obvious that Linux has already become deeply entrenched as *the* pervasive computing platform of the future.

I'm not sure Microsoft can defeat such a well-entrenched, pervasive platform.

Hope this is helpful.

F said...

Great post, John.

To me, the interesting distinction between Microsoft and Google is in monetization.

Microsoft depends on a one time sale with future revenue dependent on upgrade cycles.

Google gives Android away for free, but takes a 30% cut of all software sold on the platform. Similar to Apple but on an open basis to build scale.

How much would Microsoft be worth today if they received 30% of all software sold for Windows?

Anonymous said...

We are and always have been a Mac company and we just reinstall Windows Office on the new Macs as the old ones wear out (they do die eventually). I guess my point is that in a Linux world or a non Microsoft OS world the Office suite won't be particularly profitable either.

Anonymous said...

May be one remark regarding MS Office. A lot of governments will require the use of "open" office formats rather sooner than later. This will enable producers of alternative suites to read and process also the MS documents. Therefore, one barrier for buying these alternatives might fall in the medium term (good ol' days where MS Office couldn't even read MS Works files...).
In addition, with the interface changing that drastically, the companies will have to spend a lot of money teaching the new version. If there's a steep learning curve, they might also enjoy this with a competing product (e.g. Open Office) and have less costs afterwards. Or they switch to web-based products (e.g. Google) instead and can reduce their hardware costs.
Very enjoyable article btw.

Anonymous said...

oh, and I haven't touched the MS Media Player for years. VLC => is my preferred alternative...

Unknown said...

Based on the evidence of their current products and statements about the future from Steve Ballmer I think that Microsoft is very much on the downward slope.

Microsoft grew hugely through the 80s and 90s primarily selling software to businesses. They did this by dramatically reducing the hardware+software cost required to give a worker access to a computer. At the time they were competing with the previous generation of computer systems firms (IBM, Digital, Unisys, etc) who were selling large, very expensive combinations of hardware and software. Your article describes well how MS ran over the top of the older firms. Now however, certainly in the developed world, there is relatively little growth left in the business market.

Microsoft has achieved success by delivering software (Windows and Office) that is "good enough" for business and cheaper than the competition. It is not particularly well desgined or easy to support but it is good enough. It is in the consumer market that MS has missed the boat. The MS answer to consumer devices for the home is a PC (desktop or laptop) running Windows and Office - just like a business machine. The difference is that in the home consumer market many buyers are highly price sensitive. As you describe, because MS does not exert strong control over the hardware combinations that it can be installed on, system vendors (Dell, ASUS, HP etc) are selling cheap machines to the home consumers. Built to a price, they often have the minimum amount of memory and poor display systems. In the home environment they also are unlikely to be properly maintained which means they quickly fill up with junk software and viruses. There is no IT support team to come an re-image a machine or recover data from backups. All of these things ensure a poor quality user experience in the consumer setting.

In my opinion the biggest threat to MS is not from Apple or linux, it is probably Google or someone like them. I think the advent of a consumer computing device where the applications and data are mostly not resident on the machine but live out in the cloud is the next big step. Like on the iPhone now, most applications will have limited functionality but be relatively cheap (a few dollars) and not the major investments of the previous era (hundreds or thousands of dollars). The user hardware can be light and portable with long battery life if required. It is not a problem if the hardware breaks, gets lost or is too old. Get a new device and you again can get to your files and photos as well as things like Facebook. You don't have to transfer disks between machines. The whole idea of your digital life being tied to some particular laptop or home computer will melt away.

Mark J said...

Your basic analysis sounds right. But I think that the wide range of potential hardware in a non-Apple computer is probably more of a disadvantage for Linux than it is for Windows, at least for the time being.

A "driver" is a piece of software that interfaces between a hardware device (such as a video card or a sound card) and the operating system; without the right driver, the computer cannot use the hardware.

Most hardware manufacturers ensure that for all their hardware there is a working driver for Windows, but not all provide drivers for Linux. Thus many Linux drivers have been written by hackers themselves, perhaps reverse-engineered from what can be gleaned from the Windows drivers.

In my experience this can be a real problem with Linux laptops; you may have to wait the better part of a year until all the hardware on your flashy new laptop works (the fingerprint reader on my dell laptop still lacks a Linux driver, so it is unusable, and it's a year old now).

Of course, manufacturers will provide Linux drivers for their hardware if/when a substantial fraction of the machines they sell have Linux preinstalled.

Anonymous said...

Good commentary, mostly rings true, and I've been using Linux now for 15 years, so I've seen it come a long long way in that time.

I think one key factor supporting Microsoft has been the development platform - the tools and environment to create applications for middle tier business that led to a proliferation of software for Windows -

However now it seems that the web is it the platform, and the Windows platform no longer has the competitive edge against the competitors (such as PHP, Ruby, and Java in the enterprise).


Potato said...

One thing I don't understand about MSFT is their reluctance to move away from their forced obsolescence business model.

As you say, they finally started getting it right with XP and Office 2003, and many people saw Vista and Office 2007+ as a huge downgrade, going to huge pains to remove Vista and install XP on any new systems that came in...

Yet, they stopped selling XP and 2003 anyway, even though for XP at least, they could keep selling new licenses every time someone upgraded their hardware.

Anonymous said...

Some good posts here.

Microsoft still has a pretty firm hold of both the office and the home desktop. Its main problem is that this may not be where the action is going forward. Personally, I do not think the distinction between laptop and PC usage is that significant. Far more important is the rise of other suppliers in the IT space using wireless networks to connect to the internet. Mobiles such as the iPhone or those running Android are now allowing many users to meet their simple IT needs such as browsing emails, watching videos on Youtube, keeping upto date with social network sites etc without ever needing to touch Windows. As an example my wife has largely stopped using our Home PC since she bought an iPhone. This change in behaviour will not immediately damage the office market place but the capacity of consumer products to influence choices of business software should not be underestimated (tech geeks do not necessarily control buying decisions otherwise all of us would be using Linux already at work).

Tim said...

A few thoughts:
One of hidden gems of MSFT that no one ever seems to talk is their software development tools i.e. Visual Studio and .NET and their extensive support for third party developers. Visual Studio 2010 from the standpoint of a programmer is simply out and out better than any other software development enviornment. Plus MSFT is going to push through another forced upgrade by introducing a new graphics technology for programmers called WCF(also known as Silverlight for web-based apps) to replace WinForms. Several major companies such as Morgan Stanley and JPMorgan have announced huge internal software development based on Silverlight.
Apple has never been known for their support of third party developers and their development tools are crappy and always have been crappy. Apple I think simply just doesn't want you to use any third party software including Flash on your IPhone, thus this keeps them out of a lot enterprises that have extensive internally developed software.
Linux has okay development tools but they are all open source based which on top of having an open source OS is simply too much for many companies to bear. I suspect Linux will have a heard catching up in look and feel as WCF/Silverlight apps become much more common.

macondo said...

Operating performance - in terms of speed - is one of the arguments you mention for the advantages of Linux or Mac.

For old systms (i.e., two old ThinkPad's I own that are now running Ubuntu), I agree.

But for new systems, is performance (speed) really still a limiting factor? Other than the dreadfull start-up and shut-down times, my 3 year-old ThinkPad running Vista Business performs well with many applications running including MS Office and Photoshop.

The 'safety' issue also runs both ways - on Windows there is a whole established industry revolved around safety (and identifying lapses). Google's concerns are possibly specific to them, plus great PR.

Perhaps we could think of Microsoft being in an intermediate position between two opposites in a spectrum of service - Mac on one side (control everything), Ubuntu on the other (anything goes).

This could be a position of strength where Microsoft could simultaneously occupy various positions along the line of this spectrum - BUT, 1) it depends on quality of product (Microsoft's seeming inability to link into the mobile space is a concern); and, 2) new product execution/attitude must improve(hubris of their neighbors at Boeing seem to have infected them).

I've been thinking of MSFT too, but no idea where to start thinking of valuation. Leadership will be critical, and Microsoft first has some succession issues to deal with.

Mark J said...

You might find this post interesting ...

NickL said...

You passingly mention Microsoft's potential to shift its business model but your analysis would suggest that this may be the biggest question. Theoretically, they could create a model in which -- for instance -- the standard licensing mode is an annual subscription fee (perhaps per seat, with ability to run anywhere on any device). That may be just as lucrative for MSFT; their long upgrade cycles have meant many businesses or individuals haven't had to pay them for an OS in years.

The obvious follow up question is whether they have the capacity, particularly in management, to execute this type of business model shift. Considered this way, it seems unlikely.

Mrs. Watanabe said...

Nice post. I'm still stuck in xP, actually, and when it comes to xP (not sure with Vista or Win7), there used to be two types of licenses.

One was OEM, and you got this with, say, a Dell or other branded computer. You can't move the hard disk with an OEM licensed copy of Windows to another machine.

The other type of license was a retail license, which you could move from machine to machine. The retail license would cost much more, though, so that an OEM vendor could sell a machine with OEM Windows on it for about the same cost as a retail copy of Windows. In the long run, the retail copy was cheaper, typically, except for a laptop, which you typically can't order without an OS on it already.

Enrique said...

The buggyness of MSFT OS is clearly overstated, but a glaring ommision is the lack of support (drivers) for Linux distros. Ego may be the largest impediment for opensource OS, as geeks tend to be incapable of understanding that joe public wants simple interface, all drivers included, etc. The linux crowd will never give in to outside formatting or even logic and flow. And, how to monetize open source? Personal glory will only motivate for so long...

Kiaran Ritchie said...

Great article. It was really interesting to hear this from the perspective of a finance-minded individual as opposed to the radical left-wing Open Source programmers who usually opine on this matter.

Maybe, I'm one of those radical left-wing programmers... I wrote an article a while ago about the experience of porting my videogame to Ubuntu Linux.

There are quite a few parallels between our assessments of the state of Linux:

-The chicken/egg problem of Linux adoption vs. incentive to support it. One will never happen without the other. Hopefully critical-mass is reached soon and we pull ourselves out of this mess.

-The problem of usability and resorting to arcane shell commands. I comment how Ubuntu 10.04 was the first version of Linux I have EVER used that didn't require use of the command line (not even ONCE) to allow me to do my daily operations/setup/installation and work.

I hope you don't mind if I link to my article. I write about some of the same issues from the perspective of a game developer/programmer:

Anonymous said...

Mac OSX is partially open source. Darwin - the kernel (actual operating system, like Linux or FreeBSD) - is completely open source. Only the components built on top (user interface, applications) are closed source. There is a project called Open Darwin which provides an operating system much like Ubuntu, using Darwin as the kernel.

This means that Mac OSX doesn't neatly fit into the "closed source business model" you describe, as the underlying system which interacts with hardware is in fact open, and can be modified by anyone, just like Linux.

On the note of restricting multitasking on the iPhone, it was done primarily due to performance reasons. Having 2 applications (especially games) running on a single core 412mHz ARM processor with 128 MB memory (given the iPhone and iPhone 3G models) would be quite taxing on the hardware. Apple wanted a smooth user experience, and this is likely why they decided against multitasking. Even now with an 800mHz processor multitasking isn't "true" multitasking.

The point is that Apple didn't restrict multitasking because there are "too many configurations" to test - whether one applications runs or two, it's pretty much the same because the kernel isolates them, and it's not an issue on the desktop. In fact the iPhone does have multitasking (iPod runs in the background, telephone calls and SMSs come through, email is downloaded, etc), it simply isn't exposed for 3rd party applications.

Unknown said...

You asked for some geek comments, so here are mine:

You threw around the word 'superior' a few times. My problem with this is that superiority is about as difficult to quantify as intelligence. It really depends on your yardstick (security, flexibility, reliability, etc.)

The other thing that doesn't strike me as complete is the description of the open-source business model. Your description makes no mention of the sold-support model of companies like Canonical, or the licensing model of Red Hat, which are great examples of business models centering about open-source software. I would suggest looking into the business models of Red Hat, Novell, Canonical, etc. I would encourage you to look into what these companies sell if you haven't done so already.

The Madman said...

On the subject of installing software:

The only time you should be using the command line is when you've asked someone on the 'net for help, for example because something's broken. Then, it's just copy-pasting.

Otherwise, the Ubuntu Software Centre is an excellent tool for that kind of thing. You can just type in something vague like, "Word Processor" and have a nice new word processor installed in a few clicks.

AltansarSpiritseer said...

Great article! I only had a chance to skim through it, but will read the rest later tonight.

"Ubuntu is a distribution of linux originating in southern Africa with an explicit aim of user-friendliness"

As far as I know, this is incorrect. The NAME Ubuntu originates from Africa, that is true. The Linux distribution does not.

"Ubuntu, pronounced /ùbúntú/ (oo-BOON-too), is an ethic or humanist philosophy focusing on people's allegiances and relations with each other. The word has its origin in the Bantu languages of southern Africa. Ubuntu is seen as a classical African concept."

CKCECKC said...

You can still purchase backbone system in this country. Check out EYO.

Now I don't know what you did when you say you take the hard drive out of one machine and it flawlessly work on another. But I am pretty sure that driver support for windows is superior to Linux. Any hardware manufacturer will have to support windows - no surprises there given their market share. Can you put out a product that exclude 90% of the market? The guy next to me (I am a software engineer) just spent a whole day mucking around with drivers for his ATI card for Fedora. He is a fellow programmer so by no means a non-geek.

Your comment that with the appropriate expertise anyone can fix any bugs. I don't think that's true even for professional developers. Given someone a core file and a couple of million lines of source and you can spend says fixing it. Trust me. No one has that kind of time.

I think it's too simple a conclusion to draw that if mission critical software is build on linux then it must be solid. Windows is never meant to be a real time OS. I have seen too may linux apps seg fault and core in my life time. A lot of manufacturers are moving to linux because it doesn't need to offer a windows user experience (routers etc) and it's free to keep COGS down.

As to Google's press release about banning windows, I actually have a chuckle. I sure hope that they retain a few window boxes to system test apps like Google Earth.

One important factor to consider is that there are a large number of developers out there who have spent a large portion of their career writing for windows. In a way they have a created a ecosystem that will keep windows alive for a while yet.

If you are looking to profit from virtualisation. Probably another good place to look is boring old Intel. I think a lot of servers out there will be need to replaced their CPUs with newer i7s that have more cores. I think the energy consumption is a lot lower which should cut down on power bills - which is a very significant cost to running massive server farms.

Dulwithe said...

A great deal of mis-information in this article.

Also, and perhaps more importantly, the key to a successful operating system (and one of the main reasons Windows is able to survive so well) is the ecosystem of 3rd party software available for the OS.

For example, one of iPhone's greatest successes is the AppStore. The hardware and the operating system are good, but not mind-shattering. However, the multitudinous applications available for the iOS is what is really making it popular.

Android, also, has a good appstore that makes 3rd party software easily available to its users.

The success or failure of desktop linux will depend upon how well the linux ecosystem developers enable linux to be able to incorporate 3rd party software that INCLUDES (and doesn't avidly exculde) 3rd party commercial and proprietary FOR-PROFIT software.

Meego should be good. Let's wait and see.

Anonymous said...

Are you saying that this linux can run on a computer without windows underneath it, at all ? As in, without a boot disk, without any drivers, and without any services ?
That sounds preposterous to me.

If it were true (and I doubt it), then companies would be selling computers without a windows. This clearly is not happening, so there must be some error in your calculations. I hope you realise that windows is more than just Office ? Its a whole system that runs the computer from start to finish, and that is a very difficult thing to acheive. A lot of people dont realise this.

Microsoft just spent $9 billion and many years to create Vista, so it does not sound reasonable that some new alternative could just snap into existence overnight like that. It would take billions of dollars and a massive effort to achieve. IBM tried, and spent a huge amount of money developing OS/2 but could never keep up with Windows. Apple tried to create their own system for years, but finally gave up recently and moved to Intel and Microsoft.

Its just not possible that a freeware like the Linux could be extended to the point where it runs the entire computer fron start to finish, without using some of the more critical parts of windows. Not possible.

I think you need to re-examine your assumptions.

John Hempton said...

Are you saying that this linux can run on a computer without windows underneath it, at all ? As in, without a boot disk, without any drivers, and without any services ?
That sounds preposterous to me.


YES. It is a boot disc - and it comes with its own open-source drivers. Download it. Find an old computer and try it. It works just fine.

Anonymous said...

YES. It is a boot disc - and it comes with its own open-source drivers. Download it. Find an old computer and try it. It works just fine.


Exactly. It needs a boot disk like I thought.

Vista is far more powerful than windows XP, and runs twice as fast. It is also much harder to pirate, and this point more than anything else has the Linux crowd in a panic.

It wont be long until Windows XP is no longer supported, and when that happens, what is Linux going to do ?

Linux will have to find a way to work under Vista from here on, since it wont be able to rely on XP being readily available anymore.

Linux may seem like a good alternative to Office, but all that is happening in linux is that the windows interface is cleverly hidden away. It still needs the drivers and software services in order to run, and in most cases - that happens WITHOUT a valid windows licence.

This is just plain piracy.

Vista will finally put an end to this blatant abuse of intellectual property, and linux should decline, taking the pirates with it.

Anyone that supports the continuation of Windows XP in place of Vista surely has a hidden agenda .. and you will surely be caught out.

John Hempton said...

Ubuntu needs windows like a fish needs rock climbing equipment.


You can take out your hard drive - put a blank one in - and it will run UBUNTU. It is an operating system.

If you have not got that you have not got anything.

Half the servers in the world do not have windows on them at all.


Anonymous said...

Its clearly evident that Windows is the future. One only has to watch TV for a short period of time and see the advertising. WOW !! I personally love the part where the young man is taking a stroll in the delightful snow covered streets, and sees firsthand a young deer with a gleefull glint in its eye. It sends a shiver down my spine. WOW is all I can say. Windows is clearly the future of enterprise computing.

Anonymous said...

The OP of this comment doesn't get it. Linux does NOTneed windows to run. It does not need a "windows" boot disk, it does not run "under" windows. Why, hell, it does not even really need a boot diskof any sort because you can run it from USB, from the internet, from a networked drive, from a hard drive.

Generally, however, desktop linux is first booted from a CD/DVD as a LiveCD that will run the FULL operating system without even touching the HDD. Then, when the full linux operating system is running, the user gets the opportunity to intstall it on the hard disk with options: side by side with a previous windows or linux install, or claim 100% of the HDD.

The original poster seems to think that a windows xp boot disk is needed to run linux and that it runs "under" windows. I hope he/she reads this and learns the truth.

The Madman said...

Stupid Anonymous is stupid.

Linux, from a technical standpoint, is a kernel. The same way as under Windows, you have an NT kernel that manages hardware in order to give resources to software, Linux is a kernel that comes with its own drivers and software calls. On top of this, in most modern desktop Linux distributions, lies the GNU utilities written by the Free Software Foundation, which includes tools to allow you to boot (and manage the boot process - I.E. what starts and what doesn't), to manage files and directory structures, development tools to build software and many other tools. On top of this is the X server: a graphical server. Though there are programs available for Windows to run X-written applications, X is used as the primary graphics server of a Linux distribution and is reputable for its networking capabilities. It is often called by another program, though, such as GDM in Ubuntu: the Gnome Desktop Manager, a program that allows users to log in to a graphical environment. After the user is logged in, they come across a Gnome desktop, which includes applications for panels, desktop manager (wallpaper/icons), file management and many other tools, which is written in GTK. GTK sits on top of X, and Gnome sits on top of GTK. In Ubuntu, there are several other packages, but for simplicity's sake I'm leaving them out.

Desktop Linux is a complete and comprehensive stack, in technical terms completely removing the need for a Windows install. It doesn't use any components of Windows at all, from drivers to the GUI, and therefor DOES NOT promote the piracy of Windows. Windows Vista/Windows 7 could be the best operating systems in the world (I have to laugh at that), but that would not affect Linux development in any way. If, by some miracle, Windows was shot dead tomorrow, Linux development, use and distribution would continue completely unhindered as it is a complete and comprehensive system with no ties to Windows at all. All the code written for Linux, from the drivers to the kernel to the GNU utils stack to X to GTK to Gnome, as well as various other projects such as Nokia's Qt and KDE, are all original code: all have a copyright license that expressly permits the copying and redistribution of the software and denies the ability for e.g. large corporations to deny these rights to others, and the copying, modifying and redistribution of it has been expressly permitted by the creators.

If you thought the only way to run a computer was through Windows, you truly demonstrate the extent of your ignorance. Go back under your bridge, troll.

The Madman said...

I should also add: Desktop Linux didn't just appear overnight, like you think. Its existence didn't start with this blog post. The GNU utilities were written right through the 1980s and the Linux kernel (which as I mentioned before, manages drivers and hardware) was released to the public in 1991. Parts of Desktop Linux are three decades old.

chrispycrunch said...

A very informative write-up that basically summarizes the territory war going on among the big tech companies today.

A couple of points
- Office 2007 is a huge shift. I hated it the first time, but after you give it a chance and look at how everything is arranged, you'll find it will add to your productivity (the pop-out formatting box is genius)

- AAPL now has a market cap greater than MSFT. Do you have any comments there in your analysis of MSFT as a trade? Money flowing to AAPL stock because it's cool to own AAPL stock, means lower demand, thus lower multiple assignment, for MSFT.
- Will MSFT ever touch its pricing model for Win7 in countries like India or China? If the users decided to go legit and skip the challenges of ubuntu, MSFT could sell so many more units, at the cost of lower margins, but at a benefit of higher profits.

Anonymous said...

Agree with 狂猪's points

I think there are a few things going on here:

1) Author only recently discovered Linux and is confusing an epiphany/paradigm-shift with the novelty of discovering something new

2) Assuming that everyone has switched to laptops

3) Making the same mistake everyone else is making - writing technology articles that assume Apple is the center of the technology universe and that everything must be framed around their world

Search around a little bit, and you will find articles touting how Microsoft desktop operating system will be replaced with Linux. These articles have been running since 1995, and typically predict the demise within 5 years. Well we're on year 15 and I just don't see it.

I do technology for a living, at an extremely high level. I use Linux and UNIX on the job for most of the day for some really heavy workloads. I've used dozens of Linux distributions, and numerous flavors of UNIX including FreeBSD (which is what Mac OS is based off of), and Macs.

Out of all the desktop operating systems, Microsoft Windows 7 is just the most hassle free, and the most practical, which is why I keep coming back to it. Linux desktops have improved dramatically, however, they all still have so many quirks and lack of software I actually want to use. They are still too geeky for the average user, and I am often spending time getting the OS to work than actually running stuff on it.

On the desktop, when I want to get something done I will use Microsoft, when I want to play around I'll use Linux.

Also, Microsoft does much more than just Office, they are huge in the Enterprise, which is a big reason why the desktop works so well, since clients and servers integrate so well, and skills can be easily transferred from the desktop to a server OS.

I think once the novelty of using Linux and Virtualbox wears off the author may rethink his initial views.

Anonymous said...

@Anonymous (point #3)
Apple has a higher stock price than Microsoft. They basically run the music player business. Despite the fact that it only runs on one smartphone, they hold the 2nd place smartphone OS. Their desktop market share has been going up, compared to the Windows one that's going down.

I think it's perfectly fair to write as though Apple are the big guys in the consumer market. Right now, they are.

Anonymous said...

John, I generally agree with your posts. (Disclaimer: I am a nerd, I do build my own boxes, I have built Hackintoshes, etc.) Your linked article ends with a comment about how Android somehow shows up the "BS in the Justice Department" Antitrust Division's suit against Microsoft 10+ years ago. Excuse me? Where was Android 10 years ago? Nowhere. Microsoft's argument against the Justice Department was all about Linux and the threat that Linux posed. Here we are 10+ years later and most users do not run Linux on their hardware. I've read the court briefs in the Microsoft matter; clearly, you have not.

You grasp how easy it is to build your own desktop. Well guess what -- there are a gazillion reasonably powerful laptops available on eBay that one can buy and load with Ubuntu, or make it into Hackintoshes, and they will seem responsive to almost all end users. I don't see why more laptop sales = a continually good thing for Microsoft.

From a technical standpoint, parts of your post just irk me. You can take your hard drive from your Lenovo and get it to work in your Dell by loading the appropriate Dell drivers before you swap. Microsoft may force you to re-validate your license, which is a pain, but I have done just this type of operation and it doesn't take too long to get the hard drive running fine in the new box. And I don't mean wiping the hard drive between the swap.


General disclaimer

The content contained in this blog represents the opinions of Mr. Hempton. You should assume Mr. Hempton and his affiliates have positions in the securities discussed in this blog, and such beneficial ownership can create a conflict of interest regarding the objectivity of this blog. Statements in the blog are not guarantees of future performance and are subject to certain risks, uncertainties and other factors. Certain information in this blog concerning economic trends and performance is based on or derived from information provided by third-party sources. Mr. Hempton does not guarantee the accuracy of such information and has not independently verified the accuracy or completeness of such information or the assumptions on which such information is based. Such information may change after it is posted and Mr. Hempton is not obligated to, and may not, update it. The commentary in this blog in no way constitutes a solicitation of business, an offer of a security or a solicitation to purchase a security, or investment advice. In fact, it should not be relied upon in making investment decisions, ever. It is intended solely for the entertainment of the reader, and the author. In particular this blog is not directed for investment purposes at US Persons.