OSDev.org

Quick design in EDraw Mind Map. Thoughts?

(going a little OT here )

We haven't even said what would be the general principles of the design yet. What you came up with seems to be the typical mono kernel. So let's not rush things here. The people creating it should have a general agreement on the style first. General to more specific as we go down.

Aye.










Any thoughts?

Well personally I'd go with a more modular/microkernel approach. I don't believe everything should run in userspace though. First off, from now on what should we refer to this as, not a name maybe, but something to refer to it by(codename).

Your making a big misstake from the off, why in a market full of Mp3 player did the I-pod take off ?, because its simple design, why is google the top search engine?, because its simple design (eg: white page, logo, typing in box).
You all start off trying to compete with the big boys, your playing there game and you will lose .
Code a 3second to boot, bootable web browser and you will have many user .
You need to cut into the market, that most people use there PC for.

Who said anything about, complicated? We're going into this with first of:

Windows & Linux will die
People Will Need an OS
Our project will be in the underground

No competition involved here.

But my point is a normal desktop OS will always be bloatware, if not it will not have the things people want from a OS.
So work out what most people use there OS for and give them that, but faster
So if 70% of the time they use it for browsing the net, give them a small simple, but fast OS that you can browse on, but make it have no mean to write to hdd etc, so it stay in ram only.
Switch it off its gone, much safer, 3 second to boot.
I was thinking on the lines of us all working on F-Browser as a team .

But than again you maybe thinking you do not want to go in that direction and thats why, as a group people can not agree, we make our own OS because we like to do things differantly, in our own way.
So good luck, i wish you all well.

Dex: Ever seen ASUS ExpressGate? It does much the same thing as you suggest, and it's built on to the motherboard. The main problem is it turns a machine back into something that's not a Von Nuemann machine. The beauty of the systems we have at the moment is that they allow us to do anything we wish.

Edit: 70% I only use my machine for checking my email and web browsing. It's that 30% that I decide to compile a program or play a game that's going to be the problem. Most people don't want to learn two operating systems.

future holds a lot of secrets clearly. it's quite hard to guess what it lies for us.
the problem is that it's not 80's anymore that you can kick-start your own unix alternative from a telnet application. peeps are not interested on basic boot-loading kernel with shiny text-interface. instead the trends are being social by sharing photos and playing massive multiplayer onlines.

intel at first place was planning the moblin as a fresh kernel but plans changed, they wanted to take he power of bloated-community for their atom processors, so the choice was obvious. right now i'm playing with some django &kohana stuff for my web projects, checking XNA for an indie game but still waiting for a fresh start project to step on. i love the the kernel coding anyway

Instead of succumbing to major featuritis and then making bloatware, we need a mechanism that can bloat on itself. I mean, you just design a framework that is able to evolve (AI) at runtime with the needs of the user and then grow into a good system. Today's systems are hardly configurable.

In the short term I basically agree with Brendan. When you are asking for predictions of the future, you need to specify a timeframe. The future of OSes is an AI that directly reads/writes your brain, of course. So everything we are doing here is temporary.

A desktop does not need to be bloatware. Win 3.11 was a successful desktop with a bad API -- but by today's standards it certainly was not bloated. In the intervening years, the API has been (mostly) fixed. I would think the ultimate desktop OS would look something like Win 3.11 with a clean (all legacy support stripped) Vista API/GDI.

You are right. Configuration is VERY limited.

Especially with linux, there is no standard that distros follow.

I was going to comment on the idea of a collaborative project, and then saw this wonderful part that I can quote for emphasis: any collaborative project is going to need someone (a leader) that has the final say.

The only way you can really get around that, is by dividing responsibilities by having different people decide about different things (as long as you are careful about not getting into politics about whose jurisdiction something belongs).

I mean, it's good for multiple people to design stuff, but ultimately there are times when you either (1) need someone to say "this is what we do" or (2) you'll be stuck forever and never get anything done.

Windows Mobile 8 based on Midori will be fast powerful , solid and offer strong distributed Asynch features along with a new UI it will be a big hit.
Microsoft then release it as Windows Server X with a Windows Server 2010 VM for compatibility ( like the new XP feature in Win7) . This replaces most of their Cloud servers etc And as a file , web or managed application server it runs fast secure and rock solid.
Other OS will have a rethink to which some such as Linux cant adopt. Linux and Win7 becomes heavyweight old iron in most peoples perception.
Apple jump on the bandwagon and Release OSY with some Singularity features before MS even considers replacing Windows 7 on the desktop.
QNX will adopt Java /Mono and grow in the embedded must work market.
Web 2.0 fails and is replaced with rich internet apps. OS research ignites.

The new OS's grow and become more bloated though they are much more reliable and the kernel grows at a slow pace compared to the rest.