Steven, On Sunday 13 November 2005 21:51, Steven T. Hatton wrote:
On Sunday 13 November 2005 11:59 pm, Anders Johansson wrote:
Client/Server has absolutely nothing to do with any of this. That design concept was around before Microsoft and has nothing to do with desktops or anything else in this discussion.
I disagree. The traditional concept of client/server was desktop client/{file,database} server. Certainly there were a few others talking about client/server in the sense that Microsoft used the idea, but that was not the typical meaning intended by use of the term.
A file server is a client/server design, and was used in the Novell servers Microsoft competed with
That is exactly what Microsoft decoupled. A file server became a file service.
That was by _no means_ a Microsoft innovation. We at Locus Computing Corporation wrote the first Unix / DOS transparent file sharing software in the mid 80s. Our people wrote the INT 31 interceptor (if I've got that interrupt code right) and a user-mode, Unix-based server. The protocol was of our own design. We also wrote the first Unix-hosted MSNet file and print server a few years later. In that case, the DOS-resident software was stock the MS module but the Unix server was still user-mode and portable. It was written initially for System V Release 3 Unix (using the SVR3 TLI, not BSD sockets, though we wrote networking interface abstraction layer) and was originally developed on 3B20 desktops machines (which were _very_ nice machines for the time). I was the architect and lead programmer for that project. And to be clear, it predated Samba. Microsoft invented nothing in the field of operating systems. While I believe all the Windows operating system kernels from NT onward are good software, they are not novel.
...
Steven
Randall Schulz