How to make good use of SVN's Branch

In SVN, we generally create three folders
Trunk (trunk) Branch (branch) TAG

At the beginning of our project, we will put the code files (this does not involve document management for the time being) under the Trunk. Then we just kept working.

When will we use Branch.
According to my experience, there are two situations in which we use Branch when

    the project is stable and the entire business process and functions are completed. (Basically no bugs can be seen, only some poorly hidden bugs may still exist, such as some numerical miscalculations). At this time, we can put the Release version on Branch, and then we continue to develop new functional requirements in trunk. The relase side can only wait for some small bugs. (Testing and development done well. Infinitely close to zero).
    During the development of the project, suddenly there is a big demand. It has nothing to do with other needs. And the customer wants to see if this function can be used. So we can get a branch from the trunk side and then one or several people will work on this branch. After the customer has tested it and is satisfied with this Branche. (also mostly bug-free). Then merge into trunk.

If branch creation is not done in this way. You will find that the pressure to merge is super high. When merging, some code will be missed, and there will be many conflicts. People who merged suffer. The rest of the team suffers too.



TAG I have basically no use for myself. This is what most people say.
    After a Release is released. After a short period of modification. No bugs were found. Then we can make a branch of this Release to the Tag. Actually, I don't think it makes much sense.