GIRLS(ID:5330/gir002)Generalized Information Retrieval and Listing Systemfor Generalized Information Retrieval and Listing System Postley AIS 1961 Related languages
References: in Datamation 8(12) Dec 1962 view details in [ACM] ACM Computing Surveys (CSUR) 8(1) March 1976 view details Extract: GIRLS In 1960, Bob Hayes, a friend from INA who had offered his valuable comments on drafts of my book, urged me to join him and others in forming a new company, Advanced Information Systems (AIS), to do unspecified research and development for government and industry. I was reluctant to leave Rand but was persuaded to do so by Homer Rhodes in the course of a fascinating trip to New York. All in the course of one day, Rhodes raised my offer to join AIS, put AIS together with several other independent and unrelated companies in a new company under the name Electrada, and got the new company listed on the American Stock Exchange. Hayes and I ran AIS. He did the research (mostly government study contracts for the Air Force's Rome Air Development Center and others); I did the development. [...] Most of my development efforts at AIS were related to what I later called "File Management Systems." The initial effort was carried out under an agreement with Douglas Aircraft. Discussions with Jim Morrison of Douglas revealed that it had more than 100 large tape files developed on its IBM 701s, 704s, and 709s by a multitude of independent departments in numerous formats subject to no overall standards. Now, Douglas needed to retrieve data from all of them. I proposed to Morrison that AIS could develop a single system that would retrieve data and prepare reports from any Douglas file. Morrison agreed to fund the development. We never had a written contract; the funding came from Douglas in increments of $3,000 to $5,000, paid from time to time as needed. I think the total was about $50,000, a lot of money in those days. The result was a success. We called it Generalized Information Retrieval and Listing System (GIRLS), primarily because its principal programmers were "girls" under Dwight Buetell's supervision. Douglas used GIRLS for many years through much of the IBM 7090 era, but it was never commercially available because it was programmed in a proprietary Douglas language not available to the industry. GIRLS was a batch system. It identified its tape files by means of codes on the tapes, some of which were unbelievably strange. Then, it passed the tapes, retrieving selected information from each record and printing it in a designated format. GIRLS could retrieve any number of independent sets of data and print independent reports from each set in a single pass of the tapes. Instructions were provided to the system by punched cards created from simple paper forms defining the data and the report formats. In 1962, an internal Douglas report announcing the program's availability to Douglas personnel said (in part): This program has been developed in response to a large number and wide variety of requests for reports consisting of selected information from a magnetic tape file. These requests usually require the preparation of a new program or modification of an existing program. This program provides a more general solution to the problem of information retrieval and report generation. It combines four generalized capabilities. It can: utilize any of a wide variety of tape formats, make selections on the basis of complex criteria, produce reports in a wide variety of list-type formats, produce several reports on a single pass of a magnetic tape file. This can be done with no appreciable increase in retrieval time. By providing such a system, it is expected that a reduction in programming and machine time will be realized. These savings will, of course, be magnified for those retrieval/report generations of short production life and for those reports requiring frequent alterations in selection criteria or report format. GIRLS was described in Datamation in December 1962. As it was nearing completion, the IBM 1401with its Input-Output Control System, a first in the industrywas announced. We saw that we could retrieve files on the IBM 1401 in a manner similar to GIRLS without writing a separate retrieval program for each application. Extract: Marks I and II GIRLS was described in Datamation in December 1962. As it was nearing completion, the IBM 1401with its Input-Output Control System, a first in the industrywas announced. We saw that we could retrieve files on the IBM 1401 in a manner similar to GIRLS without writing a separate retrieval program for each application. Thus was born Mark I, a software product for the IBM 1401, so named in anticipation of the possibility that there would be a series of follow-on products on the IBM 1401 and subsequent machines. Don Sundeen, an enlisted man at Edwards Air Force Base in California, was the lead programmer, working from my flowcharts. Implementation was on an IBM 1401 at UCLA. When Mark I was completed, under contract with the City of Los Angeles, it became obvious that to be useful, it had to be married to a similarly general program to create files. I was able to set up such a project as a federally funded demonstration project, involving land-use files to be created in five cities: Tulsa, Okla.; Wichita, Kan.; Little Rock, Ark.; Fort Worth, Texas; and Denver, Colo. This system was called Mark II. After working closely with the city manager on the required data collection in a related project, Mark II was installed at Alexandria, Va., where there are a great many historical land-use parcels to track. We called it the Urban Management Data System. It worked, but it was painfully slow. Implementation of this program at Alexandria involved the establishment of databases for: transportation, public utilities, land usage, schools, hospitals, and city planning. With these files, Alexandria was able: to process a five-year census and forecast of school-age children, to develop land-use planning analysis reports, to determine suitable locations and requirements for installation of streetlights, and to evaluate the routes and the effectiveness of police patrol assignments. These tasks were done by means of requests that required information from only four separate forms to be entered into the computer from punched cards. The ability to make a request with such speed and ease, and with so little technical involvement, proved to be of great benefit to the city. In March 1965, I reported these results at a conference at UCLA in an address entitled "Economic Priorities and Public Morality" and at the annual meeting of the California Association of County Data Processors in San Diego in a talk entitled "Urban Management Data Systems." Extract: Mark III In early 1964, we were continuing to work on the faster Mark III when IBM announced the System 360. Here the light really went on when we found that the IBM 360 had an operating system. That meant that we could develop a new File Management System using the vast new functions provided by the operating system. But what about funding? The State of New York was interested in a upscale version of the land-use system we had developed and gave us a $10,000 contract to develop the specifications for such a system on the IBM 360. The system design, as with all of our file management systems, was independent of the type of data it was processing. It was designed for use by non-computer professionals to implement common data processing functions simply and quickly. Its design would enable users: easily to set up data files containing any kind of data, to update these files, selectively to retrieve specific information, to perform computations on the data retrieved, and to generate formatted reports. Since it was a batch system, it could produce an essentially unlimited number of independent reports in a single pass of the files. All of this was initiated not by writing code but by filling in paper forms designed to trigger the myriad special functions included in the system; the computer operations were controlled by the punched cards created from these forms. The report was delivered to New York State, but the hoped for follow-on development contract was not forthcoming. Later in 1964, just as work was under way on Mark III (the faster and otherwise improved version of Mark II), AIS was sold to Hughes Dynamics, a new company directly owned by Howard Hughes. Hughes Dynamics had no idea what to do with us. We continued to work on our federally funded demonstration project, staying in our existing offices. Hayes started the process of leaving to join the faculty at UCLA. It was only a few weeks later that Hughes Dynamics headquarters called me to say that it would close down AIS in two weeks. I pointed out that I did not think the federal government would take kindly to the idea that a company financed by the wealthy Hughes would leave it with an uncompleted but funded project. That had the desired effect. Hughes Dynamics headquarters called back to say that I had two weeks to get AIS acquired by another company. Walter Bauer had started a new software company, Informatics, in 1962. Bauer, a friend from the board of the Los Angeles chapter of the ACM, was happy to assume the responsibility for AIS, its contracts, and its staff. Among the factors that persuaded Informatics to make this "acquisition" was the $38,000 that Hughes Dynamics paid to Informatics to consummate the deal. AIS joined Informatics in 1964. At Informatics, we began to market, somewhat ineptly as it turned out, Mark I and Mark III as software products. Informatics management urged me to use Mark I to develop an approach to tracking commercials on television for advertising agencies, called the Media Account Control System. With the active support of Ralph Carson of the Carson/Roberts Agency in Los Angeles, we traveled all over the country making presentations to other agencies. But advertising agencies were not yet ready for software products, so business was slow. However, the Media Account Control System was successfully used to provide a service operated by the United California Bank in Los Angeles and the Harris Trust Company in Chicago, but there were no more than five customers, so, finally, the Informatics management agreed to dump the system. In 1965, we made an effort to market Mark III to Philips of Eindhoven, traveling to Eindhoven, Holland, for the purpose. That effort also did not succeed. in IEEE Annals of the History of Computing, 20(1) 1998 view details |