Informal History Of Operating System Ideas

Well, we've got an InformalHistoryOfProgrammingIdeas; what about the OS? Let's focus on new developments, paradigms, techniques, algorithms, etc. in OS design, and perhaps on major new versions of operating systems. Minutiae such as "Microsoft released Windows NT 4 Service Pack 5..." or "Linux kernel version 2.2.x which fixed bug yyy" do not belong on the list; though major revisions of Linux or Windows might...

The list doesn't have to include just major production OperatingSystems; research OperatingSystems are encouraged. Published papers, even describing features not yet implemented, are okay too.

The IBM OS/360 [IbmSystemThreeSixty](1965) was allegedly a system that pioneered a lot of multi-user/multi-tasking concepts.

One key OS idea was "TimeSharing?": one computer, multiple users, as opposed to the earlier "batch" model of just one user and just one application running from start to completion before another user/application could use the system; related to, but not identical to, MultiTasking - although terminology has varied over the years and with different systems, groups, geography, etc. "MultiTasking" and "TimeSharing?" are not single ideas, but rather complex ideas, with complicated areas of overlap.

One key figure in the history of that OS idea was JohnMcCarthy (most famous as the inventor of Lisp). He wrote "REMINISCENCES ON THE HISTORY OF TIME SHARING" in 1983, published in CACM 1992 (see http://www-formal.stanford.edu/jmc/history/timesharing/timesharing.html).

Other profitable topics to link to here:

• early GE timesharing, Dartmouth timesharing (and interaction with invention of Basic)
• invention of demand-paged virtual memory by Burroughs, and its later market popularization by IBM, and the still later virtualization-based OS by IBM (circa 1970s)
• Today's explosion of interest in virtualization technologies, ranging from Java VM to the recent VM Ware IPO.
• The notion of a HAL (Hardware Abstraction Layer), why some have viewed OS'es as nothing but HALs, and why others see OSes as more than mere HALs.
• The difference between libraries (e.g. DLLs) and kernels (users tend to not understand this critical distinction)
• Hierarchical versus older flat filesystems (also extent-based files, ISAM, KSAM, etc). Journaling filesystems.
• Multics innovations (there are many, mostly known as later reinterpreted by Unix)
• The innovation of Unix pipes as a manageable early form of dataflow communication between non-custom software components (see DougMcIlroy)
• Security innovations such as Multics security rings, Unix setuid, accounts, passwords
• The rise and fall of the idea of the "System Operator"
• Much more

---

CategoryHistory CategoryOperatingSystem