If there is one universally known and acclaimed algorithm to speed things up, it is caching. Pragmatically speaking, caching is a process of not doing the same work many times. Ideally, each distinct computational unit should be executed once. This, of course, never happens in the real world. Still, techniques to minimize repetitions by rearranging work or using saved results are very popular. They form a huge discipline named "dynamic programming."

In the context of a web server, caching usually means saving the generated response in a file so that the next time when the same request is received; it could be processed by reading this file and not computing the response again. Now please refer to the steps outlined in the first section of this chapter. For many of the real-world websites, the actual computing of the responses is not the bottleneck; transferring those responses to the slow clients is. That's why the most efficient caching happens right in the browser...