Design mystery: Some suggestions and comments for beginners in Java Design and development (IV.)

"Layering of mvc,mvc2,web design programming"
Please read the article http://forum.javaeye.com/viewtopic.php?t=11712&postdays=0&postorder=asc&start=0
"Scalable does not mean powerful, do not exaggerate its words"
The current system, because of the use of interfaces or other methods, has great scalability. But extensibility does not mean powerful.
There is an interface where users can implement their own interfaces, which is really handy. But if your system itself implements only one interface or is not implemented at all, it is not convenient for users.
For example, the validators of WebWork, itself an interface, but in fact the actual implementation of the specific classes are very small, and the function is very poor, this time if you say WebWork calibrator is very bad, Then it may not be appropriate. Of course, it is very convenient to extend the WebWork validator.
Of course, scalability is still needed, but don't brag, in this impetuous era, let us do more practical things. :)
"20/80 Principles"
At work, I often think of the 20/80 principle, the "Barredo principle". For example we can see:
Time: 20% of our time will produce 80% of the results
Product: 20% of the product brings profit of 80%
Reading: 20% of the book covers 80% of the content
Work: 20% of the work gives us 80% satisfaction
Speech: 20% of speeches have an impact on 80%
Leader: 20% of people make 80% decisions
As can be seen from the above, it is very persuasive in many cases.
Here I would like to mention a few points, but different from the above possible starting point:
1.80% of the procedures are in the handling of special circumstances, so we must attach importance to special circumstances, not because it is special circumstances, do not attach importance to. 80% of customers attach importance to special situations.
Documents are also described in detail for special cases, because developers are looking for these things when they are 80% and rarely consult documents for those that are often used.

2. Optimization problem: 80% of the bottlenecks are out of 20% of the code, so in the optimization of the code does not need to optimize all the code, only to optimize the 20% key code is enough. Of course, the pursuit of perfection of the people we will not say more.
Remember that there is an optimization principle is "do not optimize!" do not optimize, it is very reasonable.

3. If you mess up 20%, it often leads to 80% of things going wrong, or to people thinking you've done almost everything.
If you make some very imprecise comments about something, others will think you are not serious about anything else.