Back To The Future - Before the Web and the 'Cloud,' There Was Interactive 'Time-sharing'
As we watch “cloud-based” applications, Mobile UC, and IP networking take over communications between people and business process applications, I find it interesting to go back to the early days of business computers that were limited to premise-based mainframes and “batch processing” with punched cards. That’s when I got involved with enabling computers to support remote end users with keyboard terminals to access “interactive” applications.
Although most of the heavy technology lifting was done by clever engineers and software developers, I was fortunate in being able to contribute to the initial implementation of interactive (online) computing before personal computers, and now the Internet and World Wide Web, started taking over business and social communications.
Because mainframe computers were big, slow, expensive, and had limited processing power, it was not practical for individual end users to use computers the way they can now. Sponsored by ARPA’s J.C.R. Licklider, some bright people at MIT, Stanford University, etc., developed the concept of letting a number of users share a computer interactively, independently, and concurrently by time-slicing the CPU and swapping active user programs dynamically between secondary storage and main memory.
Time-sharing also introduced the beginning of what we refer to today as the “user experience” that has become the focus of good business communications. Slow input and output taking place concurrently via keyboards and printers or display terminals, allowed each user to feel that they were directly in control of their own computer application and could interact accordingly. As it still is today for online and “mobile” apps, fast response to typical user input commands, i.e., under five seconds, was a target requirement for interactive applications.
If you mention the term “time-sharing” today to most people, they will say it is “vacation ownership!” At an early 1960’s meeting of the country’s technology leaders supported by ARPA (now DARPA), I suggested that they come up with a better, more descriptive name than “time-sharing,” but the wrangling that followed didn’t accomplish anything.
Making Time-sharing Commercially Viable
Because mainframe computers were still very expensive to convert into time-sharing systems, some research organizations started looking for more commercially viable platforms to move forward with. UC Berkeley’s Project Genie was one such effort that I had a hand in helping it become successful.
Berkeley’s Project Genie chose to modify an existing, relatively inexpensive scientific computer, the SDS 930, from Scientific Data Systems (SDS) to support a time-sharing environment. However, as described by the Project Genie people, “When the system was working, Max Palevsky, founder of Scientific Data Systems, was at first not interested in selling it as a product. He thought time-sharing had no commercial demand.”
At that time, I had just joined SDS because they were starting to see the potential of developing a time-sharing product and I had been active at System Development Corp. (SDC) in the development of one of the first time-sharing systems sponsored by ARPA. I was interviewed personally by Max Palevsky and had argued with him as to the potential commercial benefits of time-sharing for business applications. He was planning to develop a next-generation business computer (Sigma) that would include online time-sharing in addition to traditional batch processing.
My ARPA connections contacted me about the Berkeley system and recommended that I look at it personally. After visiting with the Project Genie personnel and trialing its very well designed user interface and application software, I could see that it was indeed ready for the market. When I reported my suggestions to the SDC marketing management, they were not interested because of their own Sigma plans.
However, there was an upcoming computer show in Las Vegas where SDS was going to be exhibiting, so I asked the marketing manager what he was planning to highlight at the exhibit booth. He responded that he wasn’t sure yet, and asked if I had any ideas.
I then told him that UC Berkeley, an SDS customer, was giving a paper on their time-sharing system at this conference and it would be helpful to them if we let them demonstrate their system at the SDS exhibit. All that was needed was a phone line and an inexpensive Teletype machine. He thought that was a good idea and a week later reported back to me that everything was arranged.
That is when I let him know that SDS now had a problem; visitors to the exhibit will see a demo of an SDS-based computer system and will ask how much it will cost to buy one. What would be the answer?
The next day, Max Palevsky called a meeting of his management staff and decided that if anyone was interested in buying the Berkeley version of the SDS 930 computer system, the initial total cost of documenting and testing the modifications that Berkeley had made, i.e., approximately $100,000, would be added to the SDS 930 purchase price. On that basis, the answer to a buyer’s question would be “yes,” but there would be no prior public announcement.
Since I knew that all the players in the ARPA community were looking to acquire a commercially available time-sharing system, I immediately notified them all of the Berkeley system availability as an SDS product. That triggered an avalanche of orders that year and became “the most successful computer in SDS history, earning $40 million in sales and a devoted following among scientists and researchers worldwide. It ushered in the new business of commercial timesharing and was the initial hardware base for two major timesharing service companies.” (Particularly Tymshare).
Whenever I saw Bob Taylor, (Director of ARPA's Information Processing Techniques Office and founder and later manager of Xerox PARC's Computer Science Laboratory), he would always ask me, ”Art, did Max say thank you yet?”
Time-sharing technology opened the doors to real-time interactions between end users and computer applications, the hallmark of today’s online Internet and World Wide Web. It also broke down the barriers to remote access by initially using the existing wired telephone networks to provide online service access directly to end-users, eventually moving to data networks like Tymshare’s Tymnet. The bottom line for time-sharing service success was the individual end-user experience and demand for interactive applications. SDS’s success with the Berkeley system confirmed that time-to-market was a critical factor in product planning. After Xerox acquired SDS, it unsuccessfully tried to enter the time-sharing services business, but, by then, the PC had started to take over business applications.
Today, Mobile UC is expanding upon the Internet and World Wide Web in providing flexible, multimodal, person-to-person and process-to-person communications services across personalized handheld and portable endpoint devices (smartphones, tablets). UC provides an integration framework that enables end users to use their mobile multimodal devices selectively and dynamically for both personal and business needs (dual persona). With virtual and cloud-based business applications becoming more accessible to end users, there is no question that consumers and business users will all be heavily using device-independent UC in a two-way, multimodal service network environment, expanding upon the original concepts of interactive computer time-sharing.
Art Rosenberg, a veteran of the computer and communications industry, contributes his column, The Unified-View to unified communications. To read more of Art’s articles, please visit his columnist page.
Edited by Rich Steeves