Gartner Blog Network

Why does anyone care about operating systems?

by Jeffrey Mann  |  July 9, 2009  |  9 Comments

Before I was an analyst, I can remember lamenting why the people in our industry were so obsessed with chips and operating systems. It was around the time when DEC released the Alpha RISC chip, and rivalry between the different strains of UNIX and Windows was at its highest point. The horse race between BSD, Xenix, Ultrix, AIX, Solaris, HP-UX, and Unix International seemed to be a life or death struggle. Meanwhile, Microsoft wasn’t in the horse race, but was building a horseless carriage.

Chip architectures were also popular points of discussion. DEC Alpha chip (from wikipedia)At one point, I was able to sound reasonably informed about RISC architectures, multithreading and other stuff that seems pretty arcane to me now. Even then in the early 1990s, it all seemed a waste. Why so much attention to chips and operating systems? That should be far too low a level for most people to be worried about. It’s like spending hours discussing the type of nails and bricks used to build a house, while ignoring the room layout, window placement or paint color. I would think that applications and what end users see would be far more important than the details of the innards of the machines they run on.

Now, most people don’t worry about chips too much, unless you really like that sort of thing (and I am oh so glad that there are people who do, so I don’t have to). But we still seem obsessed with operating systems. Either because they are ho-hum (Windows Vista), might be less ho-hum (Windows 7), supposedly just work (MacOS X — although I disagree), or just sound really cool (Google ChromeOS).

Google has unleashed a flood of commentary and speculation by saying that it is thinking about a new PC operating system built around the Chrome browser. I will let my colleagues debate what this really means. But this all makes me a little sad. I thought that obsessing about an OS in 1993 was depressing; why are we still doing it in 2009? Next, I fear we will re-open the big-endian/little-endian compiler debate. Isn’t it much more important what we do with these operating systems? I was impressed with the thinking behind Google Wave because it shows what can be done with the clever technology under the hood. The thought of having another OS that gets in the way of what I want to do grinds down my soul like a bad third grade teacher.

I realize that operating systems and even chips are important. They make it possible to do the cool things that we can imagine. But sewers, roads, electrical grids and payment systems are important too without too much of the population having to pay too much attention to them. I will pay my share of what it costs to keep them going, but please don’t make me think about them; I have other things to do. That is how I want to think about operating systems; get out of the way and let me think about something really useful, and where I can make a difference, however small.

Category: apple  google-vendors  ibm  microsoft  technology  

Tags: chrome  operating-systems  osx  risc  unix  windows  

Jeffrey Mann
Research VP
20 years with Gartner and META Group
30 years IT industry

Jeffrey Mann is a Research VP at Gartner, covering cloud office, collaboration and social software.

Thoughts on Why does anyone care about operating systems?

  1. Zmodem says:

    I just figured I’d share with you my thoughts on this situation, so here goes 😉

    I wholeheartedly agree with everything that you’ve stated. Why are we all worrying so much about an application that drives our hardware when we don’t even worry much about what hardware we are using? I like the ‘nails in the house’ reference, there. Pretty much sums it all up :-)

    I’d like to just sit back and think where we’d be if, say, there were no operating standards. Is it possible to have a system run without rules that just engages and has applications built for it, cross-hardware, that just works? Why is the concentration here so much on bloating each bit of an operating system with some sort of competitive this-and-that for every other system out there? Whatever happened to use-ability and ease-of-use? Actually, did that ever exist?

    It seems to look as if, these days, operating system designers are looking for the most graphic intense/extreme way to ‘WOW’ the competitor’s audience rather than just saying “Our system works and it will run anything. Need we say more?” I’m tired of having to use an emulator on my Linux partition because I like a certain app that is only driven by the windows API and operating system itself. If all of these programs can be written in C/C++, ASM or another cross-platform programming language, why can’t they all be compatible with any bit of hardware that they can compile on? Why is there no universal OS that will run any application you build?

    People and businesses have spent way too much time building and catering to one set of operating system’s rules, without delving out and figuring out how it would be possible to be universal. Why can’t I insert any disc in any machine, MAC/Windows/Linux/etc, and just have it, *gasp*, work?

    When did software become more important than hardware? That’s all I’m asking…

  2. […] Google’s posted something on a blog about a technology that very few people understand and even fewer people need to care about, as Jeff Mann points out acerbically and a bit dolefully. It’s describing a fundamentally new strategy for such technology, which Ray Valdes is […]

  3. Just picking up on our conversation on twitter. ,)

    Thing is, I actually think that the OS choice is even getting more important then less. Why? Because Google, a company which is doing most of their business in the cloud, has decided to produce two new OS in two years. Why? Because their depend not only on the capability of browsers. They’ve sold that problem by supporting Firefox in a very huge way and releasing Chrome. But the ability of the browser is highly dependent on the OS and that’s why they didn’t just stop their or just release their version of a browser for the mobile platforms that already existed. No, they have built a very mobile oriented OS with Android and their stepping up on the PC market with Chrome OS which is, very similar to Android, very interlaced with Linux as OS.

    I’m really not trying to be rude, Jeff. Just thinking, that you’re wrong on that one. ,)

  4. Jeff Mann says:

    You’re right; Twitter is limiting for discussions that require more space.

    I am all for people being interested in operating systems; also browsers, microchips, power suppliers, and the quality of soldered wiring connections. Someone needs to know about them, and I’m glad that there are people out there who do.

    My point was that I’m not sure that these need to be such a general topic of interest. I had hoped that what we do with OSs would be more interesting to most people than the mechanics of the OS itself. I know that’s what I am interested in.

  5. Karl Magdsick says:

    It has always been about the performance and compatibility goblins. We have better technology, but economies of scale keep us driving coal-fired steam automobiles.

    I see a good 3 blog posts worth of stuff I was trying to cram in here, so I’ll go set up a blog elsewhere and post back here.

    People don’t really care about chips and OSes as much as they care about speed and the ability to run that program they bought 3 years ago. People who really care about elegant chips are more frustrated by how they see market forces and economies of scale railroad us into one stable local maximum, when they can see much better stable local maxima just across a shallow valley or two.

    You know how you keep telling Grandma not to click on those blinking flashing popups that say “virus detected” and link to trojan horses, and sometimes you almost want to yell at her and everyone else who keep those people in business? The CPU and OS cognoscenti feel the same way about the way we still use x86 and MS Windows.

    For instance, have you wondered why with all of its budget, and superior materials processes, Intel hasn’t gotten its Atom chips competitive with ARM in the performance-per-watt category? x86 is ugly, and if ARM and PPC (like PWRficient, not like G5) chips were more popular, economies of scale would bring us more elegant and less power-hungry laptops and other devices.

    @Zmodem : We have airplanes and we have skateboards. Why doesn’t someone just use aluminum or carbon fiber or some other lighter-than air material and make a hoverboard that just works? Come on, this is so simple! ASM isn’t cross-platform, and C/C++ can be cross-platform, but only on a source code level, and most people aren’t willing to wait for a compiler to run when they install software. Check out WINE for running Win32 apps on your x86 Linux box. Also check out LLVM for cross-platform high performance binary distribution with an ahead-of-time compiler and not forcing garbage collector or a particular object model on your code.

  6. Anon says:

    If you really want to have the OS that just work, and applications that work even when you use different computers, then maybe the “Browser OS” can meet your demand.
    With the web as the platform and browser as the “vessel” I think we can all be happy no matter what “real” OS we use. The thing is, then, you can’t have photoshop, or playing warcraft on browser.
    So? the flame war about OS or applications or programming language will always exist until we can have those “thick apps” on browser. I think this is what drives google research all this time, up until they reach “NaCl” and “ChromeOS”.
    But to think it again, the trolls and fanboism will always exist regardless technology advance. So? anyone will always care about operating systems.

  7. Dan Sholler says:

    The reason it is a big deal is precisely because it should not be such a big deal. IN order to write stuff, you need an OS. Period. In times when you know what that OS is.. then you can write stuff based on that assumption and not really worry about it. The thing is that these things usually do not change for long periods of time, and then undergo some generational upheaval.

    We went through a tremendous period of turmoil when the minicomputer came out, (anyone remember how to write code for PrimeOS? ) but that coalesced around Unix, and efforts such as Posix made it common enough not to interfere that much with software design. Windows was a given as well.

    the problem is that we are now clearly entering another period of upheaval. The operating systems we use today were designed to manage single machine instances, or at best clusters of machines that are colocated. They have all sorts of built in assumptions, not the least of which is that there is some direct tie between an application and the physical instance on which it runs.

    The OS of the future will clearly be highly virtualized, allow instances of applications to run across a network of (possibly dissimilar ) computers, that may or may not be colocated. This will require changes in the operating system, which to some extent will need to be reflected in the application design.

    (as an aside, this is why the purchase of SpringSource by VMWare may be looked upon as a significant event)

    So, you are right, in the sense that we do not want to have to worry about operating systems, because the stuff they do is too low level, but when we get to a point like now, where the assumptions on which that low level stuff is based are changing, the question of what the future operating system looks like will determine what the applications of the future should look like. This is why it matters.


  8. […] already find most Mac vs. PC discussions irritating (they both work, they both have problems, IMHO). But a split like this will inevitably encourage […]

Leave a Reply

Your email address will not be published. Required fields are marked *

Comments or opinions expressed on this blog are those of the individual contributors only, and do not necessarily represent the views of Gartner, Inc. or its management. Readers may copy and redistribute blog postings on other blogs, or otherwise for private, non-commercial or journalistic purposes, with attribution to Gartner. This content may not be used for any other purposes in any other formats or media. The content on this blog is provided on an "as-is" basis. Gartner shall not be liable for any damages whatsoever arising out of the content or use of this blog.