On Sun, Nov 04, 2007 at 01:41:37PM +0100, Anselm R. Garbe wrote: > On Sun, Nov 04, 2007 at 12:48:11PM +0100, Maarten Maathuis wrote: > > One of the problems i would see with making a completely new operating > > system is all the device drivers that have to be made or ported, that > > alone makes it a daunting challenge. 20 years ago the amount of > > devices that existed (relevant for a general purpose OS) were limited, > > there is also the question of how much will be gained from such a new > > operation system? > > Hmm, I think it is all about interfaces. If the interface is > simple and general purpose, it might be trivial to adapt device > drivers. For instance, I believe that a simple interface will > enable us to reuse existing drivers for Windows Vista or Linux, > what might be necessary is an interface stub implementation for > Windows, for Linux etc. to reuse those drivers. > > > Writing a new userland is out of the question, so you'd be strictly > > looking at a new kernel to replace either the linux or bsd kernel. > > Nah, I think if you stuck to the Unix/Linux/Plan9 userland, than > the whole idea of writing a new OS gets irrelevant. I believe > the challenge is to find a completely new and different OS > design, esp. from a userland perspective. Otherwise all design > decisions will be similiar to what Unix did decades ago. > > > I have few issues with the linux, i do not know bsd or any other unix > > like operating system. But what could be gained from the massive work > > to make a new kernel? > > Nothing, I understand Pikes paper as request for something new, > that is neither Unix nor Plan 9/Inferno (though the latter ones > come with really great general purpose concepts). I won't think > about a new OS to replace anything existing, I think about an OS > which is so different (hence new) as a demo only, that it might > show up as a promise for something in the future. It would be > enough if such an demo OS would make a specific but rather > limited thing easy to use, like a digital camera, a mobile > phone, or something else. I believe for a small community and > project, there are not enough resources to attempt replacing an > existing full-featured Windows or Unix OS.
Having said that, I think a new OS might also depend on completely new devices. Imagine that screens get replaced by headsets + eyeglasses sets or in cars with showing the UI up in the windscreen. Think of something like a mind-controlling input device for a user interface - and hence the impacts to the OS design with such things in mind... So a new OS should include some sci-fi ideas. Regards, -- Anselm R. Garbe >< http://www.suckless.org/ >< GPG key: 0D73F361
