Case studies conclusion : Why making a new desktop operating system ?

Our collection of case studies finally reaches its end, as we’ve globally described all common desktop computer OSs and even the less common Linux, from an average user’s point of view. We hence have now enough material in order to answer the following question : why should one go through the painful task of developing a new operating system, while there are other options out there ?

First of all, why not using an existing operating system, maybe improving it with my own programs ? After all, it’s not introducing a new one that will prevent me from using the others at work. But the fact is, they’re irritating me up to the point that I feel like my computer would feel smarter without them.


Windows is irritating me first because it’s bloated. On modern computers, folders shouldn’t take several seconds to open, and the OS shouldn’t take up a minute to load. 10 GB HDD requirement is insane, too, and let’s not talk about CPU and graphic cards requirements that in the end have a significant impact on laptop battery life. Then there’s the poor ergonomics : on modern releases of Windows, finding something in the control panel or the start menu becomes very hard, and exploring files sounds bad when compared with other operating systems. The “ribbon” UI that becomes more and more widespread feels incredibly messy on programs that have a lot of capabilities. Windows 7 also works poorly on touchscreens because its GUI is not able to adapt button size to a finger. Then there’s security : antiviruses are a spawn of Satan, they have a huge impact on performance and shouldn’t be necessary when the user has a fairly good understanding of what he is doing (a feeling that is often missing on Windows). And having a web browser deeply integrated in any part of the operating system is not exactly a good way of improving computer security. And then there’s the need to reboot anytime an update is installed, without being able to tell the “update” window to shut up… No, really, let’s forget Windows.


Mac OS X first irritates me because it asks me to buy a Mac. There isn’t a single Mac which satisfies my needs :

  • Mac mini : Low power, no room for a sound card that would allow me to get MIDI input for my keyboard.
  • iMac : Stupid touch mouse, LCD screen bundled with the computer (first I prefer my CRT monitor. Second I don’t want to buy a new computer anytime my screen dies), no room for a proper sound card, expensive.
  • MacPro : I’m not going to pay that much for a computer, ever.
  • MacBook : Stupid multitouch trackpad (does not know clearly the difference between scrolling and zooming), low power, no room for my sound card, I don’t want a laptop, expensive proprietary power cord..
  • MacBook Pro : Same stupid trackpad, no room for my sound card, expensive proprietary power cord, I don’t want a laptop.

Then, there’s Mac OS X itself. It lacks free software bigtime. Or more, exactly, it does not makes them run properly, because the version of the X Window Manager that is integrated is extremely poor (buggy and poorly integrated with OSX). It’s true that X is complicated and not the best way to manage a GUI, but as soon as Apple have integrated this feature, they have to make it run properly, period. And last, there’s Apple themselves. I dislike what they did with the iPod, by forcing users to use their crappy multimedia player and not allowing them to see what iTunes writes about them on the disk. But with the iPhone, the iPod touch and the iPad, they went one step further : they’re now trying to kill the concept of computing by removing the notion of freely re-programmable device. With the App Store system, Apple controls what you can do with your device, and there’s absolutely no legal way to escape it. Creepy. Do I want to give money to people doing such things ? Definitely not.


I somehow like this one. Its philosophy of freely available software and source code is interesting, and allowed extremely quick evolution. It’s able to do wonders on client-server systems and embedded devices, too. But is it ready for desktop computing ? No, no, no. To make it run, I always have to tweak things, and being not even sure that my OS will work after an update to the new release (or even an update of my graphic card’s driver) is not satisfactory. Even more, the way Linux does GUI and multimedia is hideous. There’s little to no standards, and the remaining standards are poor in some places, like the way X11 crashes and never comes back. GUI feels like an afterthought on Linux, not like a finished and integrated product (and that’s precisely what it is, indeed).

Then, there’s the core design of Linux. Made with servers in mind, it extensively uses the client-server model and is targeted to multi-users environments where a system administrator is the king. This introduces serious overhead in some places, especially when the user has to re-type its password in order to do an usual task (e.g. updating), and should be avoided on desktop operating systems where all of this isn’t needed.


Well, no one can tell if they’re going to evolve in a good way in the following years, and how long it will take. However, as of today, these operating systems are evolving… In directions that make me feel they will feel more and more awkward to me !
  • Bloat, more bloat : It seems that people maintaining Windows, Mac OS X, and Linux distributions, don’t know what to do in their days. They keep adding features, without removing a single one, but are those features useful ? Most of the time no, they sound like a way to sell the new release more that a thing of real interest. Useless features means bigger software, more bugs, more complicated ergonomics, slower loading times, more difficulty in the task of learning and understanding, etc. I prefer simplicity and cleanness.
  • Cloud computing : The latest and greatest in computer science is to jail the desktop computer user the way it’s jailed on an iPhone. Cloud computing aims at making the user rely on distant servers managed by big companies to store their data and applications. Apart from the fact that networks will undergo intense saturation and that servers will have to get insane computing power, apart from the fact that a single server or internet crash will forbid millions of users from doing anything on their computer, it’s socially speaking an extremely bad idea. If a checklist was made about what’s needed in order to establish the perfect dictatorship, it would be the first on the list, because it would allow dictators to control *anything* people do with their computers. Some people seem to forget about Google China’s strong censorship, dictated by political interests…

No, there’s no alternative to introducing a new OS in the computer world in order to improve desktop computing at the moment. Haiku sounds interesting, but I fear they may have a too strong interest at keeping compatibility with ages-old BeOS and have this prevent them from evolving. Now that I’ve defined what I don’t like in modern desktop OSs, it’s time to define what kind of OS I exactly would like to make. This will be the goal of the next part of our analysis, called “Scope statement”. Thank you for reading !

4 thoughts on “Case studies conclusion : Why making a new desktop operating system ?

  1. Emil Ahlbäck May 3, 2010 / 11:09 am

    Hi! I’m currently reading your blog from the beginning and on, and I’m enjoying the heck out of it! I’ve been a user of all operating systems you’ve discussed so far (though not all versions of them).

    I somewhat share your suspects regarding Haiku and keeping compatibility with BeOS. However I do realize that the Haiku developers can focus on the OS itself by doing this without writing a ton of applications just to keep the OS usable. Apart from BeOS compatibility I would love to hear what more you have to say about Haiku.

    Thanks for the wonderful read so far
    // Emil

  2. Hadrien May 3, 2010 / 6:05 pm

    Well, although I did not know about I by the time I wrote this article and hence just added a sentence in an edit, this is my personal favorite in today’s uncommon desktop OSs.

    I like its extremely good performance for an OS lacking hardware acceleration, and the way it helps developers getting things done with the object-oriented Kits and the translator thing (grab any image format. Treat it like a bitmap image. As simple as that). I, as a developer, also like that the team publishes guidelines on thing that matter (like user interface or general application development) and, more important, expose them clearly. Technically-wise, they’re going forward at an extremely fast pace. Look at their GSoC projects. Look at WebPositive and the OS getting a full modern web browser in a few months.

    Where I don’t like Haiku is from an OS user’s point of view, because
    -> It’s lacking on the hardware support area (that’s normal, and it’s going to improve)
    -> Its stability is poor (again, it’s going to improve. I wait for a stable release before making a definitive statement)
    -> Its icons and more generally its gui look highly dated and are often unpleasing aesthetics- and usability-wise. This needs cleanup. There’s ideas around, but…
    -> The OS bears both the BeOS and some of the UNIX legacy. At some times, they’ll have to break some of that compatibility if they want to reach the mass market. Now, when someone wants to change something in the GUI that looks clunky to the average people who comes from Windows, the BeOS crowd comes up and says “hey, no, it’s part of BeOS uniqueness”, reminding me of those unix people who can’t imagine GUI manipulation without the clunky X Windows System or prefer CLI altogether.

    I can’t state a full opinion on Haiku now. They obviously have skilled dev, so most of the issues are going to disappear. But the last one is an important one, and won’t be just automagically solved as time goes by. The Haiku team will have to break some BeOS legacy or provide some way to adapt it to us Windows/OSX people, otherwise I’m afraid that they’ll forever stay a niche OS, in spite of introducting highly interesting technology and paying attention to developers. I want to wait and see what happens. And develop my own OS in meantime instead of just hoping Haiku’s evolution will follow my needs ;)

  3. Ivan July 4, 2010 / 3:09 pm

    “first I prefer my CRT monitor.”

    mind if I ask you why? I am not trying to defend iMac or whatever, I just couldn’t think of any reason someone would prefer a CRT monitor.

    Thank you.

  4. Hadrien July 4, 2010 / 10:32 pm

    Well, even as of today, there are still several things which CRTs handle much better :
    Black rendering : When a healthy CRT is turned on, you must have good eyes or be in low light conditions to make a difference between black objects around you and the screen displaying uniform black. With a LCD, black is much more gray than it should be. That’s the beauty of subtractive color synthesis…
    Luminosity : Though LED backlight slowly start to rule the laptop and high-end TV worlds, most LCD screens still use neons as a backlighting device, which means that you can’t push brightness below a certain level. That level is already quite high compared to some light conditions you might find yourself in (e.g. my room’s 60W light bulb). Eyestrain occurs. Moreover, at this brightness, you can feel that the screen reaches its limits, and contrast is extremely crappy, making text close to unreadable
    Contrast : On a CRT, different colors are much more distinguishable from each other than on a LCD, no matter how high the announced contrast is. Moreover, as pixels are much more separated from each other on a LCD, proper font rendering requires artificial smoothing, and artificial smoothing makes things much more difficult to read than the “natural” smoothing of the CRT’s blurry pixels
    Low resolution handling : I love to play old games. Let’s face it, all LCDs suck at this. Especially with the new widescreen trend.
    Viewing angles : Seriously, why should the look of colors on the screen vary depending on where your head is ? That’s just laughable. Only high-end LCDs from NOW finally manage to address this defect, which never existed on CRTs to start with. Again, blame subtractive color synthesis…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s