Chaining or slipstreaming was done the same way in DOS as well.
What I am speaking to is the role of each wihtin one architecture. Since one is a monolithic design [no invitations to invite a debate about mach, here], and one is a micro-kernel design, it is very difficult to compare the two. We'd have to examine all three layers in Longhorn - as they work in unison on top of fundamentals in order to address how the OS will handle I/O and the FS. The thread though is about one layer, and how it will be used to deliver a better series of experiences and how these are enabled in part by WinFS. Piping one set of instructions into another isn't even close to what will be possible. Also, the development community has had and will have years to build applications that take advantage of the new OS - many, including ourselves, have begun already. I don't want to continue developing for the Unices, Linuces, or any version of them - I want to do more than what the tools available within that group will allow me to do. I want and will build a new generation of "Value Added Networks [VAN's], less the dependencies that they once had and use web services to connect business processes. I just don't see how sticking with any Unices will allow us as a smaller shop to do that. Further, I don't see any of my clients as being satisfied with one can economically produce with the tools available to the Unices or Linuces. It is not just about the interface, it is the product that processing generates - once people get a taste of visually consumable business intelligence product, they cannot go back to being flooded with data - they require product that can be used to make rapid decisions that are supported by the aggregate results of millions of dissimilar processes. As a small shop, we have to have tools to allow us to address that and the Linuces just don't - I know, we tried it and tried really hard with the best that community has to offer - I watched in horror as the companies we were/are supporting burned cash trying - they ended up unable to pay us! I mean, just because x, or y is free, does not mean that the people using it are - and trust me, these guys have the best I have ever seen. All the while, in .NET we see small groups producing thousands of very useful applications that are combined to form business systems that work for all concerned. I don't want to be restricted to > or | any more than I want to revisit the 1960's [UCSD at PT Loma], and I sure don't want to see good people and great developers fail because their development costs are through the roof and for what - the ability to bury people in data - not product.
|