Some Ideas about web applications and C/S applications

I used B/S and C/S, But I think B/S is very nonstandard, the S of B/S does not correspond to S in C/S. Search B/S or browser/server in Google. It seems that the results are both Chinese, it seems that foreigners do not often say this. I always think that what foreigners often use in this industry should be more standard, and of course it may be nonstandard; but foreigners are not very useful, it must be nonstandard. I remember a foreigner pointed to my resume and asked me what B/S meant during the previous interview. I was surprised at the time and thought: No! No, I don't know. It seems that we are wrong now.

Let's talk about things. In fact, these ideas have been around for a long time. I first came into contact with little P's cool C/S framework and was impressed with some of its cool features and wanted to implement it on the web. Later, Lao Yang suggested implementing the amendment history feature to start thinking about it. After careful consideration, I found that web implementation is not as simple as in the C/S architecture. Consider the simplest C/S architecture: client-DB server, corresponding Web architecture: browser-client (Web Server)-DB server. In a web application, browser to client (Web Server) is a stateless link, which is the problem. To implement some functions in C/S, web overhead or cost is relatively high. For example, to implement Amendment history, the web server must maintain the current state of the data before sending data to the browser (avoid using the word "client" here). Simply put it into the session, the browser sends the next update request and then compares the data. Compared with the implementation in C/S, it is certainly a lot of trouble. Besides, State maintenance cannot be implemented simply by session. We cannot put a large amount of data (or objects) into the session, and in a cross-domain or distributed environment, session synchronization is also a problem.

View csla. net, I was very happy to see its undoable (Change UNDO) feature, and thought it was also supported by the web version of the Framework. After reading it, I found that it was only supported by its c/s version. It seems that the stateless nature of Web determines that some implementations in C/S cannot be copied.