OSDev.org

I'm thinking in the terms of Windows vs Linux vs Mac.

For Windows the user base is the mass public. Generally people don't care how it works as long as it works. This caused the problem with IE6 hanging around so long...
The programmer culture is generally about the commercial side. A software package, game, business app etc is created knowing it will reach the largest audience. There isn't a lot of pushing the boundries coding happening on the Windows front.

But going back further to the DOS days... The user base was generally business or gamers, not many mum's or dad's using it for general entertainment like they use a PC now. The culture was more the home hacker, maybe people who moved from the 8 bit home machines to the 16 bit PC. They could poke around with DOS and control the hardware in ways the OS didn't bother with. I also see this kind of thing in the Amiga culture.

Linux is more the hacker space now. The user base is business, experimental tech developers or tech educated home users. The coding culture again is the hacker where they can take advantage of the system.

I don't know the Mac demographics well so my idea of this might be wrong. The Mac users are generally as technical as a windows user but they know better than to use windows... I couldn't tell you anything about the coder culture.

My target would be going back to the DOS/Amiga days. I want to target people who wanted to get down to the hardware level with coding without having to control every device themselves. Provide only the base system in a protected way. I see the culture being self taught hackers, the system will be simple to understand with only very little hidden details.

So who's your target user base and what's the coder culture you would want to build?

My target audience is me and if I am lucky anyone who likes what I do. I guess in general tech-savvy people who likes alternative solutions to things operating systems have been able to do for decades already.

What kind of solutions? What's an example?

My target is ganeral public and espacially, Computer scientists.

I disagree.
Linux has already taken over the server and it's on its way to winning over desktop users.

In my opinion it just needs better 3D graphics support (most games run like crap, even on a high end card) and then I will be able to delete my Windows partition; I only use the Windows partition for playing games now.

Weird. Most of my games run very well on my GNU/Linux box:

Yeah the server market would be the business end and the desktop is the tech educated users I was talking about. I don't see it ever being as simple to the average user as Windows. There's a weird mix of Windows users from DIY PC builds that linux doesn't have driver support for, gamers who have the best variety of games compared to linux or mac and then just the regular folk who already know windows and wont be bothered to ever switch. Hell I got family that wont upgrade to a smart phone cos it's more than they need to make a phone call.

Maybe one day linux will own the desktop but I can't see it in the immediate future.



Yeah dunno how this could be fixed except to get more game developers to get more gamers to get more game developers...

Honestly I haven't played with linux for years as a single desktop for everything, so maybe it's changed more than I can imagine.

As a side note I'm not a fan of Linus, maybe he made unix popular but there are cleaner, less restricted (license wise) versions of it.

Why not Zoidberg (sorry FreeBSD)?

It's nice to know that games that look like they come from the DOS era run well on your Linux box. Meanwhile, in the real world....

Here's a crazy idea, maybe it takes a gamer to make a gamer OS...

Do Guitar Hero III and Minecraft count? Neither runs sensibly on anything that doesn't have more than one core.

As I understand, you want to create a base for low-level coding. People, who want low-level programming, just develop their own systems and don't need any bases, but maybe I am wrong.

It's not only Linux' problem. For example, Nvidia doesn't provide a lot of documentation, the company also is going to disallow unsigned ROMs for their cards. When you play 3D games with the open source nouveau driver (pre-installed on most distributions) you really can see a lot of lags. The only solution is using the proprietary driver by Nvidia, but it also has some limitations.

My target groups are
a) users who dgaf about what works how and who just want to download & start their apps
b) programmers who want to develop for target group a)
(main goals: usability, simplicity, performance)

This kind of "eco system" requires a little more care on the developer side, but much ease on the user side.

Low level base and coding every device is 2 different things. Being able to write to a frame buffer is not the same as having to write directly to the svga registers or use a not quite right for pmode vbe 2 interface (vbe 3 as far as I know isn't widely supported).

This applies to the ata, keyboard, sound card etc... Just knowing you can program almost directly to hardware doesn't mean you have to.

And the last but I think the most important point

OS coders are not Device Driver writers...

We want to implement things on top of hardware. We want to test schedulers, task swapping, paging, process environments and threading. If that means it's slower by a few ms to access the frame buffer, disk drive or usb controller then that's fine by me until I do write native hardware support.

This sounds a lot like windows already