Editing JOSS (section)

==History==
===Initial idea===
In 1959, [[Willis Ware]] wrote a RAND memo on the topic of computing in which he stated future computers would have "a multiplicity of personal input-output stations, so that many people can interact with the machine at the same time."{{sfn|Marks|1971|p=1}} The memo gained the interest of the [[US Air Force]], Rand's primary sponsors, and in 1960, they formed the Information Processor Project to explore this concept, what would soon be known as [[time-sharing]]. The project was not specifically about time-sharing, but aimed to improve human-computer interaction overall. The idea at the time was that constant interaction between the user and the computer in a back-and-forth manner would make such interactions more natural.{{sfn|Marks|1971|p=1}} As JOSS director [[Keith Uncapher]] later put it:

{{Blockquote|There were two perceptions that I think drove the experiment. One was, from at least 1950, when I joined Rand, there was always a focus in the [[computer science]] part of Rand toward smooth user interaction. It was just built into us. We learned it from Gunning. ... all we thought about. The other was, seeing the mathematicians struggling with Marchant calculators, not being able to use JOHNNIAC in its raw form. It was just too hard. They had to learn to be programmers. It was just a frustrating experience for them. We wanted something that would look to them as being a natural way of thinking about an aid to the solution of their problems without very much training, and without any perception of becoming a programmer, because that would have kept them away.<ref>{{cite interview |last=Norberg |first=Arthur |date=10 July 1989 |title=An Interview with Keith Uncapher |url=https://conservancy.umn.edu/bitstream/handle/11299/107692/oh174ku.pdf |page=7}}</ref>}}

A formal proposal to develop what became JOSS on the [[JOHNNIAC]] computer was accepted in March 1961.{{sfn|Marks|1971|p=1}}

===JOSS-1===
JOSS was implemented almost entirely by [[Cliff Shaw|J. Clifford Shaw]], a mathematician who worked in Rand's growing computing division. It was written in a symbolic [[assembly language]] called EasyFox (E and F in the US military's then phonetic alphabet), also developed by Shaw.{{sfn|Ware|2008|p=62}}

The JOSS system was brought up formally for the first time in May 1963, supporting five consoles, one in the machine room and another four in offices around the building.{{sfn|Marks|1971|p=iii}} The early consoles were based in the IBM Model 868 Transmitting Typewriter, as the Selectric had not yet been introduced to market when development began.{{sfn|Shaw|1965|p=3}} The first schedule was published on 17 June, with JOSS running for three hours from 9am to 12 every day.{{sfn|Marks|1971|p=2}} It was declared fully operational on eight terminals in January 1964.{{sfn|Bryan|1966|p=2}} The final version was deployed in January 1965.<ref>{{cite book
 | url=http://www.rand.org/pubs/research_memoranda/RM5270.html
 | title=JOSS: Central Processing Routines | publisher = RAND
 | first =JW | last = Smith
 | id=RM 5270 PR
 | date=August 1967
 | access-date=2012-04-16 | type = reference user guide}}</ref>

By this time the JOHNNIAC was already over a decade old, and its tube-based logic was never highly reliable to begin with. Even when it was working well, the system became so popular it quickly bogged down. Users were enthusiastic, one stated:

{{quote|People adjust their lives to fit around JOSS… No use coming into RAND before 10:00 am when JOSS arrives, in fact noon or after 5:00 pm is a better time, JOSS is less busy. When JOSS starts typing answers, the titillating pleasure is equaled only by the ensuing anguish when JOSS breaks off into {{sic|jibberish}} or goes away commending your code to oblivion. We can hardly live with JOSS, but we can’t live without it.{{sfn|Shaw|1965|p=14}}}}

Another put it more succinctly:

{{quote|It's better than beer — we're hooked.{{sfn|Bryan|1966|p=2}}}}

===JOSS-2 switchover===
In May 1964, the decision was made to look for a new machine to replace the JOHNNIAC and dedicate it entirely to running an expanded version of JOSS. The machine would also have to support new terminals made to Rand's specifications, and be delivered by 31 October 1965.{{sfn|Marks|1971|p=2}}

A total of nine bids were received for the new machine. [[Digital Equipment Corporation]] (DEC) won the contest with their new [[PDP-6]] system, and Air Force funding was released for the purchase. DEC also agreed to build thirty terminals based on the [[IBM Selectric typewriter]] modified with a special mechanism to advance to the next page in a fan-fold paper feed. Several other portions of the overall system were delivered from other companies.{{sfn|Marks|1971|p=3}}

The PDP-6 arrived in late July 1965, and was initially tested using [[Teletype Model 33]]'s as terminals. The new version of the code was developed by Charles L. Baker, Joseph W. Smith, Irwin D. Greenwald, and G. Edward Bryan. The system was first declared operational in October, although this included six hours of scheduled maintenance per week.{{sfn|Marks|1971|p=9}} The first prototype terminal arrived in November. In December, a terminal in [[Las Vegas]] was connected to the machine remotely for the first time.{{sfn|Marks|1971|p=3}} In February 1966, this was sent to [[McClellan Air Force Base]], followed by one in August to the [[Air Force Academy]] and two in September to [[Advanced Research Projects Agency|ARPA]] and an Air Force office in [[the Pentagon]]. The first permanent offsite teletype connection for a [[Teletype Model 35]] was installed at the [[Langley Air Force Base]] in February 1967.{{sfn|Marks|1971|p=9}}

With the new machine up and running, JOHNNIAC was taken offline on 11 February 1966, and officially retired on 18 February. Its last running program was written in JOSS and counted down seconds until it would be turned off. The machine was sent to the [[Natural History Museum of Los Angeles County|Los Angeles County Museum]],{{sfn|Marks|1971|p=4}} and eventually ended up at the [[Computer History Museum]] outside [[San Francisco]].<ref>{{cite web |url=https://www.computerhistory.org/revolution/birth-of-the-computer/4/94 |title=JOHNNIAC |website=Computer History Museum}}</ref> Having been replaced by the JOSS-2 system, the original was retroactively known as JOSS-1.{{sfn|Ware|2008|p=15, 109}}

===JOSS-2 use===
By the end of 1966 the new JOSS-2 system was fully functional and turned over to JOSS use 24/7. The new machine offered about 30 times the computational speed, five times the storage space per user, and many new features in the language itself.{{sfn|Marks|1971|p=5}}{{sfn|Bryan|1966|p=2}} The new platform could ultimately support up to 100 terminals in simultaneous use.{{sfn|Bryan|1966|p=3}} The Air Force remained the owner of the system, with Rand and others operating as consultants. [[CPU time]] was billed out to external users at a rate of about $1 per minute, although that was only during the actual execution; time spent typing and printing was free.{{sfn|Marks|1971|p=32}}

By 1970, there were between 500 and 600 users at Rand and various Air Force sites across the country.{{sfn|Marks|1971|p=10}} Many of these were occasional users, precisely what the system had been intended for. To support them, the custom terminals were equipped with special "JOSS plugs" so they could be wheeled from office to office and plugged into custom outlets. Two-hundred of the plugs were installed around Rand, and the terminal chassis was designed to be narrow enough to fit through doors while still leaving room for a coffee on one side.{{sfn|Marks|1971|p=15}} Toggling the power switch on the terminal caused it to connect at the new location.{{sfn|Bryan|1966|p=3}}

The custom terminals had been designed for JOSS as the engineers felt that the Model 33's "telecommunications functions are a mystery to the untrained person" and would be too confusing for the occasional users to bother learning.{{sfn|Marks|1971|p=14}} In practice, the Selectric-based mechanisms proved too fragile for the constant use they saw and were frequently being repaired in the field by IBM servicemen. In 1970, a full third of the terminals were completely rebuilt, and the output from the computer slowed to improve reliability. The Air Force, by this time, had decided to use the Model 35 instead, as it was both readily available and much less expensive.{{sfn|Bryan|1966|p=12}}{{sfn|Marks|1971|p=9}}

Uses were enthusiastic and to serve them, Rand began publishing ''The JOSS Newsletter'', edited by Shirley Marks. Ultimately 44 issues were published between November 1967 and June 1971.{{sfn|Marks|1971|p=47}}

===JOSS-3===
In the early 1970s, programmers at one of [[IBM]]'s west coast facilities built a JOSS-3 for the [[IBM System/370|IBM 370/158]] running [[OS/360]]. Little information about this system exists, with the exception that it was on this machine that JOSS finally retired.{{sfn|Ware|2008|p=65}} According to a note in a historical overview, Rand was hesitant to allow IBM to use the JOSS name, and as a result "the IBM version was never widely used."{{sfn|Ware|2008|p=16}}