The Slow Death of the PC: Still Waiting for a Viable Challenger

Rob Enderle

The PC is clearly going through changes. Apple is surging and spent the last several years distancing itself from the idea (and name) of PC. Its coming iPad is more appliance-like and more dependent on Web service. While netbooks kind of bounced on a similar concept and now are more like notebooks, the coming wave of smartbooks suggests another try. Eric Schmidt, who I'm convinced is trying to destroy Google, popped up suggesting that the PC was dead in three years. While I agree with the direction, and actually wrote on the death of the PC model and the Apple-driven side of this, the timeline and comment reminded me more of how often we are wrong when predicting change like this.


Let's chat about why the PC is in decline but why it will likely remain a powerful player for the rest of this decade.


It comes down to the fact that, except for some bleeding-edge types, we don't much like change.


Death of the Mainframe: Circa 1980s


I was at IBM at the time that IBM itself seemed to declare the death of the mainframe era. Analysts and companies alike claimed that the mainframe was dead, to be replaced by client/server computing. Around two and a half decades later, the mainframe is still around and it is still one of the largest and most profitable businesses IBM has. Far from being dead, it is now a showcase for some of the most cutting-edge technology density, energy management, and cooling technology in the market. While client/server went the way of a heavy focus on price, eliminating much of the R&D budget for revolutionary advancement, the mainframe's strong margins allowed for changes that make the mainframe of today arguably more technologically advanced than its newer client/server counterparts.


Some of the most important and mission-critical applications in the world still run on mainframes, which remain at the top of reliability and availability performance metrics in the areas where they are in use. Given how profitable this is for IBM, it is very likely that a decade from now mainframes will remain both relevant and powerful in the market. They aren't dominant and won't be, but they aren't dead, and won't be, either. In a way, in a world transitioning to appliance computing, mainframes may be the ultimate appliance.


We often forget that change comes hard. The more complex the ecosystem, the harder that change is to make. The mainframe ecosystem was one of the most complex. As a result, despite our predictions, the mainframe is arguably healthier now than it was at the end of the '80s.


Fax Machines


Back in the 1990s, the market moved to e-mail. It was clear that the death of the fax machine was imminent. But we're still talking about it today.


You can scan documents and pictures, in color, and transfer them electronically faster and more efficiently over e-mail. Fax machines aren't secure and bypass electronic document controls, making them ideal for intellectual property theft; a simple mistake can easily send otherwise confidential documents to the wrong fax machine with no ability to recall them. (Granted, recall doesn't work that well with e-mail, either).


Nearly two decades later, fax machines remain common, though they're clearly declining, and while they may be built into printer/scanners, they generally are favored for written document exchange, particularly where signatures are required. Their rate of decline is glacial. Back when I was forecasting e-mail deployments, they were also the key technology holding back e-mail because, for many, they were good enough.

Death of the PC Is Premature


There is no doubt that the PC is changing. A bunch of us put forth our visions for the future for Windows and while there was some disagreement as to whether Microsoft could change, there was no disagreement that it needed to change this platform for it to remain relevant. However, when you realize that the vast majority of the installed base is back on Windows XP and will remain there for at least two or three years, you'll realize that this is not a market that moves quickly. We measure its moves in decades. Eric Schmidt is likely off by a factor of 10 in terms of predicting the death of the platform. However, that doesn't mean that the PC isn't in decline. Or that, like the mainframe, it will need to go through massive changes to survive.


Developer interest is clearly shifting to things like the iPad, iPhone, Android, and even the new Windows Series 7 platforms. The father of the modern PC, IBM, left the market and the parts suppliers are vertically integrating. Microsoft, at least for the consumer segment in their stores and smartphones, is setting hardware specifications and changing who actually owns the customer. This suggests the PC is attempting to evolve. Like the mainframe, it may not be dead in our lifetimes. The question of relevance remains in play and clearly the iPad, smartbooks, and Google will have a lot to say about how relevant the PC is in the future. Still, the PC won't go without a fight and we really don't like change that much.


Wrapping Up: The PC Is Waiting


Coincidentally, while I was finishing this, I got an e-mail from one of my oldest friends on a conversation between Iran Air Traffic control and a USMC FA-18 fighter pilot that went like this:


(Note: this could be a version of the warship lighthouse story.)


Iranian Air Defense Radar: 'Unknown aircraft you are in Iranian airspace. Identify yourself.'
Aircraft: 'This is a United States aircraft. I am in Iraqi airspace.'
Iranian Air Defense Radar: 'You are in Iranian airspace. If you do not depart our airspace we will launch interceptor aircraft!'
Aircraft: 'This is a United States Marine Corps FA-18 fighter. Send 'em up, I'll wait!'
Iranian Air Defense Radar: (no response .... total silence)


Suddenly, I want to toast a Marine. However, given the limited success of alternatives to the PC to date, the USMC pilot's response it so on point. Before we can forecast the death of the PC, there needs to be a real challenger. Until then, like that USMC pilot, the PC won't be going anyplace.

Add Comment      Leave a comment on this blog post
Mar 8, 2010 8:13 AM a. asdf a. asdf  says:

Cloud computing started out as a benign concept and it's beginning to turn into a very dangerous one.

The PC hardware revolution of the past several decades was based on consumer needs for faster and cheaper computers to power their local applications. Now as applications are moving to the cloud, consumers are no longer looking for better and faster hardware.  When (if) the next killer app that requires advanced hardware comes along, consumers wont have the hardware to make it work.

Mar 8, 2010 8:42 AM Rob Enderle Rob Enderle  says: in response to a. asdf

Good point, if OnNow works (and we should know in about 3 months) that may not matter because that will showcase that performance can be pushed the the cloud as well.   Many think it won't work so this is a "wait and see" thing.  

Mar 8, 2010 9:26 AM a. asdf a. asdf  says: in response to Rob Enderle

What is OnNow? Links please (yes, I've Googled it, even Binged it, nothing)

They are not going to be able to push hardware intensive apps to the cloud. I don't see how they can ration out a 3Ghz processor and 3GB of memory to a million customers at the same time, unless there is some kind of technology I am unaware of.

Mar 8, 2010 11:49 AM Rob Enderle Rob Enderle  says: in response to a. asdf

Try this is full gaming performance (Crysis level) hosted.  Service assigns the performance you need based on the game.  Flat rate.  Launches in a few months.  Steve Pearlman is CEO.  Seen the demos and beta has gone well.   Does require a good DSL or cable connection (latency more than bandwidth).  But if he can do this and scale it...   Secondary game revenue through demos (you can basically demo anything full resolution).  Uses either a PC browser, or a very thin client.  Cool huh?

Mar 9, 2010 1:23 AM a. asdf a. asdf  says: in response to Rob Enderle

OMG! THAT THING HAS TO BE STOPPED! I'm here trying to convince people cloud apps are bad, they got entire game consoles in the cloud. And you giving statements like "So they either embrace it or it's a go-out-of-business " is not helping. (lol)

Seriously, I don't think what they are trying to do is technologically possible. I'm sure a whole bunch of people smarter than I have already figured it out, but I don't see it. Have you ever tried to play an online video with zero buffering? The reason you need buffering is for when you have late or missing packets, you need a 'buffer' zone to make up for it. With OnLive, you have 0 buffer, everything has to be streamed as soon as possible. So you're going to see every late and missing packet. And I have no idea how they've solved the hardware issue. If a standard PS3 can serve 2 players at the same time (I don't know how), they're going to need n/2 PS3's for their subscriber base. That's a lot of hardware. If they do pull it off, can you put in a good word for me during IPO time.

Mar 9, 2010 1:50 AM Rob Enderle Rob Enderle  says: in response to a. asdf

Beta has gone well, we'll see if it scales.   It only works on low latency networks and within 100 miles of the hub.  If this works it will change gaming forever.  If you can get high performance apps to work in the cloud, why wouldn't it be better?  You should see the demo. 

Mar 9, 2010 7:24 AM a. asdf a. asdf  says: in response to Rob Enderle

I think I forgot to put a lol at the end of my all caps paragraph above, lol.

The 100 mile hub is interesting. Ultimately, you could end up with OnLive server farms for each community where people just rent boxes and games and connect through a thin client. At that point who cares about changing gaming, you've changed computing. If the experience is good enough, just offload your Windows boxes and everything else up there. If I can play games off the network, I can pretty much do everything else. And yes that would be better.

Have you tried this service off-site? Do they have plans for hosting OS's etc?

Mar 9, 2010 7:27 AM Rob Enderle Rob Enderle  says: in response to a. asdf

Not allowed to talk about their plans yet but your imagination is clearly on the right track.   Yes have seen this off-site, no question it can work.  Question is whether it can scale, looks great on paper.  We'll see how it handles large loads.   It is really a fascinating model.  

Mar 10, 2010 7:22 AM a. asdf a. asdf  says: in response to Rob Enderle

I've been watching the OnLive demo on Youtube;

I'm still trying to figure out what the catch is. He touches upon things like buffering and lost packets, which was the first thing I thought about when I saw the service. According to the video, the physical limit is 1000 miles and 80ms of latency. He talks about giving the micro consoles away and such, but nothing in terms of future development. He's basically holding a full blown network computer in his hand. The kind that Ellison would have only dreamed of back in the 90's.

Mar 10, 2010 9:10 AM Rob Enderle Rob Enderle  says: in response to a. asdf

Yes sorry, 1000 miles, forgot a zero.   Should be able to go farther once those new Cisco central office routers are in.  Actually the future development stuff he is keeping close to the vest.  It is really fascinating though.   The catch will be server loading.  No really good models on graphics loading of servers yet, just what they got during the beta and for a service, beta isn't a great way to test full load and models are unreliable when something is as new as this is.   Should be interesting to watch this roll out later this year.   It is cool though huh?

Mar 10, 2010 9:39 AM a. asdf a. asdf  says: in response to Rob Enderle

Yes, very cool (roll out date is 6/17) How does one even start a service like that? I thought my tech ideas were crazy. Imagine going to a bunch of investors and explaining how you'll be streaming Crysis over the net to homes and playing them on a micro consoles. Going to game publishers and trying to convince them their games can run on consoles the size of a matchbox.


Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.