Summary of small items related to development and OLAP

development and OLAP related small projects, there is a question that we must clarify. It is also a very important one, that is, what the customer actually needs?
what exactly do they want us to develop? What are the functions? This is a user requirement analysis that is often emphasized in system analysis.
it must be in place. However, as a system analyst, after learning about the user's needs analysis, there is another important problem: technical implementation.
, although the specific implementation problems may not need to be considered by full-time system analysts, most system analysts in China's software industry still need to participate in specific programming work, therefore, this requires our system analysts to consider whether the functions required by the customer are implemented.
there is a conversion process, it is to take into account what the customer seems to be doing during programming, and when the customer uses the system, they do not
the operation is required, instead, let's take a look at a result of computer processing.

Recently I developedOLAPFor a small project with a little relationship, it was first met by other staff of the company. They analyzed the requirements and when I came here
I only read some of the documents I wrote at the time. I feel that there are many places for user operations, and many dimensions are required to be left to the customer. But the more I do, the harder I feel.
Because we useDundasThe user also needs to be able to perform some statistical work, so you need to override the control or modify the control or useGDI +Come
Writing a display control is too troublesome.

 

After more than a day of thinking, I think these jobs can be well counted before they are presented, because these statistical jobs are actually based on databases,
That isOLAPIt doesn't matter much. After all, the user wants to see the results rather than conduct data mining. Therefore, the task becomes simpler.
However, if we want to do data mining, We need to write many things by ourselves. It is not ideal to use general display controls. Many users want personalized statistics.
No work can be done.

 

I think we can make more conclusions when developing some small projects, and some findings will be useful for future development, rather than simply using them.
Users may not describe the requirements due to various factors, or their descriptions may mislead us, which will lead us
It may take some detours at the beginning.