Thursday, October 17, 2013

Windows as a Legacy System

I have been seeing charts like this one all over the place, but I think this one is telling.  It's from a presentation by a firm called KPCB.

First there's this technology cycles graphic:

Finally, here's the visual chart that goes with the above infographic:

So the WinTel era starts with MS-DOS in about 1982, and if you plot the line at the left side, with some kind of long tail, a gaussian distribution, which I think makes sense, you get this:

If that happens, here's what you'll see:

  • PC shipments will continue to decline, and the PC will become a tiny niche that represents about 10% or less of the overall computing world's annual sales.
  • Microsoft will continue to control the WinTel/PC desktop world, but the significance of that will shift almost entirely upstream to the Enterprise and large Corporate markets.

Yesterday, when I was having trouble sleeping, I did a search on a few Legacy Technologies that you might not remember.  Does anyone remember the HP e3000 family of computers, and its accompanying MPE/iX operating system?

Imagine your business ran on these boxes:

Then imagine you got a letter from HP saying that they're ceasing production of this hardware you rely on, and the software that runs your entire business on it.  They politely suggest you transition onto Something Else.   Your compilers are no more good. Your source code is no good.  Your tools and your skills are no good.

Okay enough fear mongering.    Such a thing is very unlikely to happen in the WinTel PC world, at least not before 2099, which is long after I've left this planet, and you too, probably.    

But we're creeping ever closer to the era in which Windows applications on a Windows PC are a niche object, something not everyone wants anymore.  Businesses will continue to use them until there are no more bits of hardware left that run them.  But the days of PCs getting twice as fast are already hard up against the nanoscale physical limits of the universe, and the days when the R&D budgets that drive this Moore's Law growth are going away too.

Soon, very soon, within ten years, Windows will be a shadow of its former self.  This, my friends, is why Microsoft is scared, and is reacting with Windows 8.0 and 8.1. You and me, ordinary developers who target Windows are going to be fine, just fine.    In 2200, there will probably still be virtual machines running some of the Windows software that you and I are building now.


  1. The chart presented IMHO skews the reader's sense of what is happening. They present each category as a percentage of total number of devices. Fine, that view has a place. However, it does not show a decline in the number of desktop devices, rather it shows that desktop devices become a declining percentage of the total.

    There are very many applications which will remain on some sort of desktop device for a very long time. Would anyone wish to edit a complex spreadsheet on an iPhone? Or edit a feature film or television show on an Android? For that matter, consider the security implications to corporations of their most valuable data being viewable on a mobile device, possibly away from the secured network....

    Yes, the total number of "computing platforms" is on the rise. But many of these platforms are compromised to their form factors, just as laptops have always been. I have a 17" screen laptop, and I would not, as a matter of choice, use it for development. The keyboard has a poor feel, I do not like touchpads, and although the 17" screen is nice, I am by now quite addicted to my dual 27" screens.

    To put the Kleiner Perkins data into perspective, it would be nice to see another chart, where the Y axis is number of units, and then to present the market share data in that context. I have no doubt that the K-P data are correct, but the interpretation of those data is the issue.

    If you create some number of mobile apps which go viral, you won't care about the desktop. But few of us will do that. For many of us--perhaps for most--the desktop will remain a bread and butter environment for a long time to come.

  2. It's definitely skewed, but what we have currently is a mature healthy Desktop Computing World. It's stable, it's mature, and the potential for huge new profits is slim, thus the incentive for future investment on the desktop is also slim. The pace will slow, and Desktop will become, as we all turn gray and retire, the new Mainframe/Cobol/LegacySystems space.

  3. Sure, soon server will run on phones and watches and glasses.... - as many others applications really used in production systems. There are too many fools around who read IT yellow papers and dream about a future that is not what you think....
    And many of you fails to notice that many new devices are used together a Windows PC. It is not like in the past when a new PC or OS displaced the old one. Now I can use a smartphone, a tablet and a PC all together. Just IT "journalist" who can barely use one device at a time, has to show they can use the right one...

  4. Thats all nice.. but in my company we will not run the ERP on android tablets.. maybe for salesstaff CRM but not inhouse...and that will not change for the next 10 years ...

  5. What I think is that people that were misusing PC's are now finding viable alternatives for their day-to-day tasks. I also see that I like to check my calendar on my mobile. I just think that the PC's will be once again less crowded. gamers going back to consoles, managers/ marketeers going to mobile tablets, kids to tablets, programmers going to all devices !

  6. I don't think it's that simple. I have spoken to several other Developers who find that the demand is shifting to Web Apps wherever possible. that means, that the PC is no longer The Platform anymore, The Web is the new dominant platform. Anything that can run a Chrome or Firefox level browser is now the platform, for 90% of business apps.

    Seriously. You are thinking that iPads are replacing PCs. That's not really what I'm saying. I'm saying that the layers and boundaries and goalposts, and the playing field and the rules of the sport of Football, as you know them, are going away. Hold on Tight Dorothy, because Kansas is going Bye-Bye.

    1. And where web server runs? On PCs.... where are webapps designed and developed? On PCs... Where trading, especially HFT, is performed? On PCs... what controls production at many automated sites? PCs.... Or what makes medical equipement work? PCs...
      And beware: not everything can run in a web browser yet. Try to design (maybe in 3D), try to perform some sophisticated calculations... and so on... the business world in not made only of ERPs and CRMs web clients. Where is your 90% from? Look around, and you'll see the "mobile" hype is pushed by media because they have their own interest in it - especially because it is the perfect *consumer* device. Look around you, and you'll discover PCs are still everywhere, and will be for a long time.
      Moreover you're contradicting yourself: mobile devices are killing web apps too. People don't want a web app on their device - they prefer a native one. Why Google delivers a native Youtube app on Android while forbidding everybody else to develop one?

    2. You segue from "several developers" to "90% of business apps", a figure of which I would be highly skeptical. Although they have become much more capable, browsers are still blunt instruments, and fit the old maxim that when you have only a hammer, everything begins to resemble a nail.

      Wish all you like, but true business apps are going to stay on the desktop for a long time yet. Things I would hate to attempt from a mobile device (whether in browser or not):
      - serious spreadsheet work
      - serious documentation
      - non-trivial graphics manipulations
      - any sort of CAD

      One reason the desktop will remain is productivity. The keyboard is not going away any time soon, because for most people on computers, it remains highly effective. Large screens are also not going away. Look at how many people, and even businesses, are spending serious money on larger displays these days. They do it because such displays are effective.

  7. What I think changed with this map is what "computing devices" mean. Because the more computing devices are the smaller the margin is. If you have 200$ to invest per month in IT for your family, you will likely pay less for software to provide enough hardware so as an user you may have to look for lower cost games (or free ones) on mobile phones. In fact the graph is certainly screwed as printers are not shown, consoles, etc. shows like 350 million PC will be sold this year. Comparing with consoles (in 2012 were sold at over 140 millions of PS3 and XBox), the graph should show at least some other OSes publicly are not represented. If we count car computers (that will become smart too, if they are not already, some TVs, etc) the graph is even more screwed.