Washington State Association of Data Processing Managers Newsletter banner

Volume 9 No. 11 November 1979


December 6, 1979


Bob Gillespie
Vice Provost for Computing, U of W

"Computing Services at the University of Washington"
Golden Carriage

Speaker Introduction

One of the major State data processing units is housed at the University of Washington. Computing services on an academic campus share many of the problems faced by State agencies; they also face unique challenges.

Bob Gillespie, Vice Provost for Computing at the University of Washington, is familiar with both the problems and the challenges. At the December meeting Bob will compare and contrast data processing management on the academic and state campuses. He will describe the data processing resources of the University and discuss their use. In particular, Bob will consider the impact of the University's mission and a changing computer environment on data processing management.

Bob Gillespie denies having the answers to all of the questions facing data processing managers. The information exchange, however, promises to be lively on the 6th. This is one lunch you should not miss.

Structured System Design Technology

We began this series of articles by identifying two basic problems with traditional system design; lack of a good methodology, and lack of ability to interact with people. We then reviewed the various design methodologies and chose the structure chart technique of J. D. Warnier as the one to refine to meet our needs. We then discussed the ideal design technology and the emphasis on the productivity of people.

The current methodology we are now using could well be labeled Warnier-Orr-Brackett. It was originally proposed by J. D. Warnier in his book, "Logcal Construction of Programs." Ken Orr used the basic idea of data charts and logic charts, but proposed the development of code directly from the logic chart. We took those charts and developed a complete methodology for designing total systems with direct user involvement.

Our methodology is shown in he chart below, which itself is a logic chart. Any enhancement, be it new development or upgrade, begins with a definition of what is to be done. This definition begins with a statement of the problems and/or needs, which runs from one sentence to three paragraphs. A longer statement tends to repeat or over elaborate and detracts from the focus on the problem or need.

Next, a set of objectives is formed which is the criteria the new or upgraded application must meet. As these objectives are formed they are written in complete simple, understandable sentences. Negative objectives are allowed, i.e., items that will not be accomplished. This doesn't mean that all global non-accomplishments are listed, but key items that are discussed and are chosen not to be included may be listed for clarification.

Constraints may be listed, but generally the reciprocal of the constraint is listed as an objective. This allows a more positive approach to the situation since generally objectives are positive and constraints are negative.

Assumptions may also be listed, but again the reciprocal of many apparent assumptions may be stated as objectives. If assumptions are stated about the design, they must be resolved and either stated as objectives or removed from the list prior to any physical development. One of the worst things that can be done, either in the traditional or the structured approach, is to move into implementation with pending design assumptions.

The development of the problem statement and objectives is done with direct user involvement, usually on a board in a conference room. When an objective is agreed to, and written on the board, it is an implied sign-off~ although objectives may be modified during this definition step. However, when the session is concluded, data processing and the user have agreed to the statements. They can simply be typed and become part of the documentation. There are no memos from one party to another expressing one s interpretation of a meeting and asking the other party to sign-off on that interpretation. This is one of the key points of our methodology. The user and data processing work cooperatively to state the needs and define the objectives to meet those needs. There are no memos, no design narratives to sign-off, no hardware or software constraints to the needs and objectives. The communications are kept open and constructive.

In fact, we find in most cases the user inherently knows the solution to their own problems. They generally lack the experience on how to state the solution, and those of us in data processing sure haven't done much to help the user state the solution in the past.

It's much like a person that wants to build a house (the user) who comes to the architect (data processing) for design assistance. It's that person's house that they have to live in. They should have a direct say in what it's like and how it functions.

The same is true in data processing. It's the users system, and the users data. We should design what they want, not what we think they want. There are, of course, time limits and dollar limits, etc., but it is basically the user's choice and our job is to assist them. So many of the system problems today have arisen because data processing assumed what the user wanted, or gave them a little more than what they wanted, or arbitrarily gave them less than what they wanted, or just plain badgered them into accepting what data processing could provide.

In a recent class I stressed direct user involvement and no user sign-offs. One persistent attendee argued strongly in favor of user sign-offs, because his boss required it. I finally agreed but stated that contracts were two way. If we required firm user sign-off, users should require a firm sign-off from us as to due date and cost. This fellow literally exploded out of his chair, threw his arms in the air, and said, "____ ____, we can't do that".

The only alternative is direct user involvement and open communications. There should not be any long design narratives, and no user sign-offs. Users do know their jobs better than we do, and together we can design very efficient and very effective systems. The first step is a definition of the user's need, and in many cases the best solution is a good definition of the need.

Mike Brackett
Department of Fisheries

Structured System Development Design Sequence

An industrial psychologist stated that the number one problem of business and industry is boredom. But, he adds, it isn't necessarily the repetition in a job that creates dissatisfaction. Many highly complicated and sophisticated jobs are repetitive. Rather, he says, it is the resentment and depression that workers feel when th y are denied the opportunity to use their skills.

Bits & Pieces

Association Biennium Dues

The following agencies and associated voting members have not paid their biennial dues:

The invoices were mailed to division chiefs on July 1, 1979. If additional invoices or other assistance to obtain payment authorization are required, please let me know,

Patti Palmer

Sharing helps avoid redundant effort

Association Minutes - November 1979

The Association of Data Processing Managers was called to order by President Mike Brackett.

Mary Jo Lavin introduced Gordon Sandison, Director of the Department of Fisheries, as our guest speaker. Mary Jo indicated Gordon's past affiliation with the State Senate and the Higher Education Board.

Gordon began with a little history of our state. Salmon was a food staple with the Indians as well as the other settlers. Settlements were always located on the mouth of streams. Early treaties between Indians and settlers were all verbal.

Fishing fleets have always been popular even during the Depression. In fact, during that time 600 persons made their living catching fish; now more than 4,000 persons make their living in that way. During the Depression no one fished the ocean.

Indians continued using "terminal" fishing methods: Indians would not go after the fish but wait until the fish came to them, Later fish became an item of commerce as well as food for the Indians.

In the 1960's, the Indians began large net fishing. Recently a method called "gill-netting" has become popular, We have 3,500 boats, The nets used in gill-netting extend 150 to 200 yards and float with the current, Fish try to swim through the nets and their gills get caught. This results in the capturing of many fish.

Large trollers have been on the waters for three years. Generally 30 to 35 feet with one to two persons manning two lines with 20 lures each. Now the trollers have 8 to 9 lines with 50 to 60 lures each, These large trollers also contain freezing systems right on the boat allowing them to stay out for weeks at a time.

Indian tribes have been experiencing arrests for fishing out of season, The tribes want their fair share of the fish, In District Court, Judge Bolt interpreted the treaty as "a fair share of the fish to the Indians is equal to 50 percent of all fish caught;" plus, they are allowed to have all the fish they need for food and ceremonial purposes. Judge Bolt allowed no appeals, The Department of Fisheries was unable to manage the fishing resources and commercial fishermen felt abused, The Indians in turn abused the subsistence and ceremonial catches. The Governor and Mr. Sandison went to the U.S. Supreme Court to get a ruling. In the meantime, the Washington State Supreme Court ruled that the Department of Fisheries had no jurisdiction over settling disputes of fish assignments between races, On July 1, 1977, Sandison was found in contempt of court for his inability to insure the Indians 50 percent of the catch, The Supreme Court ruled the 50-50 split was to be in existence; however, the subsistence and ceremonial fish must come out of the Indian's share. Also, that non-Indians must take the fish that they take home out of their fair share.

Tribes and the Department of Fisheries are working well together now. The tribes have the best hatchery sites which are located on reservations. The U.S. Government has pumped 33 million dollars into the hatchery improvements. Billy Frank, of the Nisqually Tribe, is a genius in the biology area. He has fish plants and hatcheries that are' very successful.

On the coast a different problem exists: 1) All natural stock has been lost because the nests were exposed to sun, seagulls etc. Result? A very poor fishing year. 2) The Department of Fisheries used to control fish for 12 miles out from our coast. Now we only control 3 miles. The other nine miles are controlled by the U.S. Pacific Fishing Council. Solution? Last year ocean fishing was closed and today we have the best spawning ever.

Still there are too many people chasing too few fish. The gill- netters and the trollers are catching enormous amounts of fish. Charters are becoming very popular. Twenty years ago there were 16 charters and they took around 5 people per boat. There are 376 licensed charters now each carrying around 20 people. Last year the charters had a very bad year with no gas and with the publicized bad fishing year, people were reluctant to go as far as Neah Bay. The Pacific Council also put a limit of two fish per person for all ocean fishing. Therefore, most charters are undergoing an economic catharsis. Many want out.

Another problem with the natural stock is a heavy siltation which buries the gravel and the eggs. A new machine from the University of Washington is being put into effect to clean the gravel.

A controversy is underway between the United States and Canada to determine fishing rights. The Canadians used to fish off our coast and our commercial fishermen sued them and they lost their rights. The Canadians are now insisting that the people in the U.S. may not fish in the Frazer River. Our dilemma, according to Mr. Sandison, is when the fish come out and turn right, we are unable to pursue them.

Gordon Sandison's closing remarks specified, "a need to clean up the salmon habitat and find out what he is up to."

There was no Treasurer's report.

Jim Hichal announced that the Data Processing Authority meeting will be held Wednesday, November 7 in the House Office Building.

Mike Brackett and Dick Applestone met with Leonard Nord and then with Tim Seth to discuss the manager's matrix. A detailed report will be present at the December ADPM meeting.

The Deloitte, Haskins and Sells testimony was handled by Mike Brackett. The thrust was productivity of people. Roberta Giovannini has a copy of the testimony.

The meeting was adjourned.

Association of Data Processing Managers Meeting Agenda

Golden Carriage December 6, 1979 12:00 Noon

  1. Introduction of Guests
  2. Presentation of Guest Speaker
    Bob Gillespie
    Vice Provost for Computing, U of W
  3. Treasurer's Report
  4. DPA Announcements
  5. Old Business
    - Job Matrix - Dick Applestone
  6. Correspondence
  7. Comments from Members
  8. Adjourn

Bob Gillespie will discuss computing services at the University of Washington - a look at some of the present and future challenges for data processing management.

REMINDER: Please send your notification of attendance to Patti Palmer, Hail Stop KF-01.