Optimization is the continuous improvement in the existing design

Intermediary transaction http://www.aliyun.com/zixun/aggregation/6858.html ">seo diagnose Taobao guest cloud host technology Hall

The first user gets information through the network.

In the beginning, the information on the network is relatively small, then the user only need to remember a few sites on it. And with the increase of the content of the network, the information is more and more large, the user obtains the necessary informations, the cost is more and more high. At this time, technology changes, the emergence of search engines, improve the user's access to information efficiency, which is, using a better algorithm to improve the efficiency of execution. When faced with searching for more personal information, demand has changed, and traditional search engines are not. A new technology emerges, human flesh searches.

Second, in terms of system deployment.

At the beginning of the information age, the software used by the enterprise is still small, the initial cost is lower, the maintenance cost is very low, so the software is generally bought back to install and maintain. However, when users use more and more software systems, the initial cost and maintenance costs are more and more high, the enterprise load more and more, this time need to optimize. Cloud computing is timed to bring together the software users use to maintain the center, while users pay for the use of the software. This pattern is very similar to Linux under the process, in Linux process code segment memory is shared, data segments and so on each process one. We can use the enterprise analogy as a process, the cloud computing analogy to the various processes common code segment, in order to improve efficiency.

Third, why software performance is getting lower.

At the beginning of the design, the performance of the program may be good, but when the requirements are increasing, code changes, the performance of the program is more and more slow. There are several reasons for this, when demand changes, we are not to consider how the software logically more reasonable, but how to change the existing code more convenient, which led to the code as the requirements change, the logic of the program increasingly unreasonable, resulting in an offset.

In addition, when demand changes, some cases are no longer running, and we tend not to delete the corresponding logic because it can lead to errors. For change in demand, we prefer to add something to it instead of reducing it. This leads to more and more code and a lot of useless code. The software is the same in our day-to-day processes. For example, a bug management system, one day a leader asked to count the data, when he asked the program to add a field to the bug. But after a while, the leader does not want this data, often programmers continue to enter more than one field.

The improvement of efficiency may not require much work.

Or in terms of bug management, generally we are through the way the site to manage bugs, every programmer needs to constantly refresh the site to see if there are new bugs. Such a programmer's quick response to bugs depends on how often they refresh the bug system page. If its half day query once, then it may waste half a day of time. But if we let the bug site get a bug out of the programmer, send an email to the programmer, assume that the user's message is open in real time, and check the mailbox every 5 minutes, then we can make sure that the programmer's response time to the bug is shortened to 5 minutes. It's easy to add Mail on a website, but it's likely to be more efficient on a large scale. The improvement of efficiency is not necessarily proportional to the effor cost of optimization, the key is to find the right place.