well back to the original article it makes sense that ibm is dumping money into linux. it makes sense for big companies that have deep pockets where economies of scale can dictate better returns for putting your code monkeys to work. gnu/linux will most likely be cobbled together for large "enterprise" situations to get rid of the microsoft tax.
it is interesting to read the arguments against UNIX because of the systems i am familiar with i prefer unix-based systems. i can see your arguments that unix is archaic/dated, but to me that signifies a well-designed system; something was designed to be elegant, but not overly-simplistic. there are a host of other systems that haven't had the staying power of unix. vendors only have so much say but more realistically are just their end users' lackeys.
it doesn't matter what system you write code for they all have trade-offs. for instance BeOS that was mentioned above had some really interesting ideas with threading, db file systems, i/o, etc. but they were just neat ideas. a pet project. it won't be integrated into the mainline systems because it didn't capture mindshare. i still think Unix has the upper hand as it has become canon for most techies. the reason is that it doesn't suffer from needless abstractions. everything is a file. the network is the computer. and so on. at some point it is just a stream of bytes this way or the next with the ability to perform computations and change the streams. why muck it up any more than it already is? look at microsofts products. they went the non-unix route and it has served them humble pie at every turn.
but i guess my whole argument is it is all contingent upon developer mindshare and leveraging your developers/users time most effectively. the nuts and bolts of why one system is superior over another really is a waste of time and energy. if you can figure out how to get X,Y,Z done in less time because you can leverage an existing code base or developer's knowledge then your are that much farther ahead of your competition. it makes sense to not have to re-invent the wheel if the wheel works but just needs some tweaking...
Unix will become the one true way(tm) and if you don't like it, well tough. you can always build your own kernel, own system libs, applications, and support files if you don't like it. but oh yeah that's right, that won't happen. you will leverage other people's coding just the same as the rest of us.
It boils down to this- how much time do you spend on your jig vs. end product? You will never have a perfect jig. But if you get it close enough...
-jaromski
it is interesting to read the arguments against UNIX because of the systems i am familiar with i prefer unix-based systems. i can see your arguments that unix is archaic/dated, but to me that signifies a well-designed system; something was designed to be elegant, but not overly-simplistic. there are a host of other systems that haven't had the staying power of unix. vendors only have so much say but more realistically are just their end users' lackeys.
it doesn't matter what system you write code for they all have trade-offs. for instance BeOS that was mentioned above had some really interesting ideas with threading, db file systems, i/o, etc. but they were just neat ideas. a pet project. it won't be integrated into the mainline systems because it didn't capture mindshare. i still think Unix has the upper hand as it has become canon for most techies. the reason is that it doesn't suffer from needless abstractions. everything is a file. the network is the computer. and so on. at some point it is just a stream of bytes this way or the next with the ability to perform computations and change the streams. why muck it up any more than it already is? look at microsofts products. they went the non-unix route and it has served them humble pie at every turn.
but i guess my whole argument is it is all contingent upon developer mindshare and leveraging your developers/users time most effectively. the nuts and bolts of why one system is superior over another really is a waste of time and energy. if you can figure out how to get X,Y,Z done in less time because you can leverage an existing code base or developer's knowledge then your are that much farther ahead of your competition. it makes sense to not have to re-invent the wheel if the wheel works but just needs some tweaking...
Unix will become the one true way(tm) and if you don't like it, well tough. you can always build your own kernel, own system libs, applications, and support files if you don't like it. but oh yeah that's right, that won't happen. you will leverage other people's coding just the same as the rest of us.
It boils down to this- how much time do you spend on your jig vs. end product? You will never have a perfect jig. But if you get it close enough...
-jaromski
