Rapid web development under the. NET Framework (I)-tangled miles

I have been engaged in software development for five or six years. I wrote it on my blog.ArticleRelatively small. There are many technical elites who love programming as well as me, so I decided to write my enterprise development experience so that I can discuss, learn, and improve it with everyone.

I will officially use it.. NET development also takes more than five years, from. NET 2003 to the current 2010 along the way, because the company's projects are relatively complicated, so do not focus on. net for in-depth study, generally the company refers to where to hit, I believe that corporate development friends also have the same experience with me.

The first project is to develop an enterprise conference system (OA ), it involves mainstream modules such as "document Customization", "workflow", "to-do work reminder", "permission set", "user", and "organizational unit. Due to the tight schedule, the project entered the coding phase from the very beginning, and the research was conducted at the same time. There were not enough people. I remember there were four people in the project team, one responsible for researching and writing documents, the other responsible for project management, and the other responsible for me.ProgramMember. In the development phase, the research is conducted at the same time, so there is no corresponding business requirement report. Because some functional modules have been identified in advance, the development starts at the same time. Overtime .. .. Rework .. And so on. The process is not detailed. This is my first project, but the result is satisfactory. The four project teams completed the OA system within four months. I remember that the most difficult thing to do at that time was document definition, especially when it involved master-slave form data. My biggest achievement through this project is that developing a quick platform for business expansion can greatly improve efficiency and greatly facilitate business changes. For example, the addition and subtraction fields can be easily changed through the platform, and there is basically no need to encode the fields, so there is no need to compile and deploy this series of processes.

Then I entered another project, which lasted more than two years and has not been fully accepted yet. Of course there are many reasons. I will not discuss other reasons here. I will only talk about some experiences in the development process. This project has a semi-mature platform for development. When I first came into contact with this platform, I was sure it was good. The idea of implementing many functions is similar to that of oa I have done. The difference is that the concepts such as data objects are introduced, that is, the object tables with a relationship are abstracted into an object, and the fields in the table are abstracted into the attributes of the object. This platform has two major modules: a development platform and a design platform. The design platform is relatively simple, that is, the investigator adds data items and the results can be exported to SQL to generate data tables. The development platform abstracts the entities defined by the design platform into an object. The early stage of this platform provided us with convenience. Soon the software prototype came out, and the basic module came out. However, there were a lot of problems with the change of project functions. For example, there is a clone function in the platform, that is, to copy the module quickly, and the problem is that in the early stage, it was fast because of it, and later the problem was also caused by it. The replication module usually uses the same data object. If the two modules use different data objects later, the underlying changes of the entire module will occur. It is equivalent to re-starting from data objects, pages, transactions, verification, etc., with a huge workload. The customer feels a small change. We have been busy for a long time and haven't solved it yet, so the construction period is dragged down. There are also some defects of the platform, because there was no source code or complete help documentation at the time and it was not solved. So far, this project has not been fully accepted. This project makes me realize: (1) project management is the top priority. Because at least four project managers have been replaced, the coupling problem between project modules has not been solved, this has paved the way for future maintenance of the project. (2) Management: In addition to reasonable personnel transfer, supporting documents and plans, project management must provide guidance. (3) for management, the project risk assessment must be done well. For example, the use of third-party platforms and plug-ins should fully assess whether the project is controllable and the risk size. This is what project managers should do. Of course, it is best to develop your own R & D when time permits (4) Fast platform development brings speed while, the high coupling brings great risks to the project.

Of course, there is no rainbow without wind and rain. Only through these tangled Miles can we make progress. We will continue to discuss fast web development on the. NET platform and hope to discuss and learn together.