ABSTRACT
Many friends ask what made me decided to write the blog series The Beauty of Mathematics in the Google China Official Blog. Although the original purposes that the company published official blogs were to introduce Google's technologies and products directly to Chinese users, I always believed that users would like your products better when they understood the background technologies behind feature they could see. As a research scientist who used mathematics in my daily work at Google, I tried to explain the mathematical foundation of Google products in a language that most people could easily understand. I also liked to should young engineers how to use mathematics they learned in college in their engineering works. In the US as well as in China, I often see that the majority of software engineers rely on their intuitive feelings to navigate unknown territories, solving problems using makeshift methods. This problem is especially pronounced in China. Such an approach is highly unsystematic and unprofessional. When I first joined Google, I noticed that some early algorithms (such as spellcheck) had no systematic algorithms or theoretical foundations, but were entirely made up of makeshift phrases or doubles of words. Although these methods are better than nothing, they have no potential to progress or improve. Instead, they make the program's logic rather chaotic. As the company expanded and grew, Google started to hire engineers with excellent theoretical backgrounds from the world's best universities, which ensured engineering projects’ correctness. In the following year, I mentored several graduate students from some of the world's best universities. Using a hidden Markov model framework, we unified all of Google's spellcheck models. During those years, Google rewrote almost all of the projects’ programs, and there were no more makeshift methods. However, in other companies, including some second-rate IT companies that are ostensibly titled high-tech, bad practices were still prevalent. In China, some small startup companies sacrifice quality for quantity, which is understandable. However, makeshift approaches persist in some listed companies with profits that rank high in the world, giving people a rather poor impression of them. On the other hand, these companies often surpass many transnational corporations in their spendings on office buildings and management offices. This is like the nouveau-riche who dresses up fancy but neglects to make progress in knowledge and character. Hence, I wrote The Beauty of Mathematics in Computer Science in the hope that these companies’ engineering supervisors can lead their teams to improve the engineering quality and move away from poor practices. Only this way can they approach the level 256of world-class IT companies and avoid astonishing waste of resources due to large amounts of low-level repetitive constructions.
