The Evolution of Operating Systems – Per Brinch Hansen (2000)

Overall I found the paper to be quite informative especially as I had not been formally introduced to the history of operating systems from as far back as the 1950s. Early on in the writing and references, it is easy to see the characteristics of the Open Shop systems and the Batch processing systems as a “one computer for all” type of architecture and of particular interest for me was the massive scale of the expense (dollars) in terms of wasted computer time even though each user had only around 15 minutes to execute their programs. I could then see how the Atlas Supervisor, being called the earliest operating system made sense, as it was the first of its kind to explore concepts such as spooling, demand paging and supervisor calls, which we can see clearly exhibited in the operation of modern operating systems today. Most of us are familiar with the term spooling most commonly used in terms of printing in general or with a few other output devices that are much slower than our modern computers today.

The concept of Timesharing which was proposed by Mc-Carthy in 1962 was also key in leading to much of our modern execution of multi-tasking and it was interesting to see how this developed from the demonstration of CTSS into the larger Multics File System at MIT which although it was not a huge success in its entirety, gave rise to the first hierarchical file system for both private and shared files. Unix, having succeeded Multics in terms of time sharing systems, has become widely adopted, even today, I have personally worked with some variances of the Unix-cloned operating systems and flavors of Enterprise level Linux (such as CentOS). I also found the database problems highlighted by Stonebraker (1981) to be interesting and I intend to do more research into these for my own personal knowledge.