Demand control?
Kingfish ---- 2004/8/16
Http://www.niufish.com/archives/2004/000161.html
These days are all in nineteen, and the diapers have changed again. The critical moment of the diapers will also play a role. I have been called "Baba," by my colleagues. I suddenly thought of it when I worked, the demand for software change ...
First Of All, if the customer requests to change! Do not change, it is garbage, change is software.
Rule of thumb, most of the needs come from customers. The changes in demand are not existent. If you do not decompose the software into a number of phases, from the demand phase, the design phase to the encoding, test, the integrated demand change is included in each stage. Why is this?
Customers have their own demand! That's right, but why? Which methods do we use to solve these problems?
because:
1. Customers do not know what they want to do, the customer knows their needs, but there is a misunderstanding when communicating with software personnel. I thought the software personnel understand, but it was too late when I saw something I imagined. 3. The original customer's clear demand is just a thick line, and when the software personnel makes the thin lines, it is not right.
We are so resolved:
1, CMM: Refining demand refines all demand in the demand phase, let customers confirm. If the demand changes in the future, the operational demand change process is recognized by the SCCB (version change committee) (large demand change). 2, RUP: It is gradually refined to the last stage that continues until the last stage. Participate in the iterations of each part to ensure demand is gradually close to the goal. 3, XP: Demand can be changed at any time, using user card tracking requirements. Use object-oriented methods to control change costs.
Is it solved?
1, cmm: This guy's solution is very "should". If the demand is not detailed, it cannot be guided. If the demand changes, the cost of change is increased, and the sccb can decide which to modify it, which does not change. It seems to be a lot of specific problems in reality: 1.1, refining demand is just idealized ideas, which will directly be affected by three reasons unkreated by customers. 1.2, Software personnel in SCCB tell customers if the costs required to change demand changes, and customers will make a decision. But this organization contains many people, project managers, developers, managers, etc. The developer will say that we need a lot of time, the project manager will say that we haven't time, the management will say that we have several projects that cannot be added, and customers generally have only compromise for project progress. 1.3, CMM has a process method of college gas is very loved by Chinese administrators and thinks this better "control". In fact, the customer is unknown whether the needs are unknown, and will be changed through the demand change process. This seems to have to recognize your own mistakes, and the software personnel pushed all the responsibilities to the customer. No one is not a mistake, just like renovation you can say that the designer will definitely guide your sample? Are you not changed once a time? 1.4, do you carefully read 2/3 of the user needs instructions? When I encounter such a multi-page thing, I don't want to see it. You think about it before the exam, this book is all focused. This will definitely cause the appreciation of understanding. If you meet the communication document, this will cause waste. 2, RUP: Demand is refined, from thick lines to fine lines. This method determines the direction of the demand and has little effect on the cost of the project when the demand changes. It looks very good, but the resource cost of communication needs is also very big. 3, XP depends on the extent to which object orientation, emphasizes the importance of communication.
Why do I think of ninety-nine:
