I think that, for you, a Gentoo style distro might be the right thing.Seems like others here are a bit afraid of this. no offense but what gave linux the chance to stand up against windows in the first place was the reason that many people could work on it. not everybody knows the same amount about networking, multimedia, kernelling and perhaps some other hardcore stuff. the idea there is that the people who are proficient with a certain topic take on the driver (or Server in BeOS terms) and put it to perfection (as far as possible). i do not see a problem in this way of thinnking. the only problem with linux is that (and BeOS shares somewhat the same problem) it is based on c++ (or c) which is the source of many, many problems with improper programming. writing a program in c++ is one thing, making it stable another but making it really good is a masterpiece.
This is exactly the reason that we are working things the way that we are - there is no "one size fits all". We can make a distro that is nearly perfect for any given user. That is pretty easy. The hard part is making a distro for EVERY user. That is the hard part. And that is why we aren't doing that - we are letting others pick niches. Someone wants to be the Gentoo of OBOS? Great. Another person wants to be the RedHat of OBOS? Great, too.
Mine too. :-)Yeah... you can't get past him in a serious study ^_^
Well, yes and no. I mean - BeOS provides many (most?) of the same services as Windows or Linux. I think that if you calculated the number of bugs per feature, you would find that BeOS did a whole lot better than Windows. :-) I certainly don't intend to "bloat" BeOS. Small, sleek and fun is where it is at.i refered in this one more about the driver load and compatibility load. what makes windows a pain in the ass to use and makes linux sometimes a pain to config is the attempt to satisfy a large amount of devices out there. i mean i still cannot use my networking card i especially bought once to work under linux, windows and beos. somehow the driver doesn't recognize the card although the chip should match the drivers. and without network connection an os in todays world is doomed.
Linux isn't exactly where we want to be, but it is far closer than where Windows is. I want to be at a place where end users can, 90% of the time, only install "big" releases - that is, R1, R2, etc. That they don't need patches. That patches are available for certain weird cases (i.e. we didn't test on a brand X motherboard with Y meg of memory and there is a weird interaction that we fixed in a patch). I don't want to have to release a patch that says "Prevents any website from rooting the box". That is a *BAD* bug. I hope to not have those via good testing and many eyes.c++ is the source of evil... procedural programming in this special case. most of those bugs are buffer overflows or underflows which result in a wrong usage of memory. as long as you rely on c++ as your main os language you will always confront those problems. a good framework can reduce the problem but not extinct it.