I deleted the original post on MO and put it here.
Unlike the western counterpart, Chinese family names have fewer varieties. So you can see many Chinese mathematicians named as Zhang, Wang, Li, etc. In the past, this seems not to be a big problem in the math community since some famous Chinese mathematicians happen to have an uncommon family name such as Yau and Tao. But as there are more and more famous Chinese mathematicians with common family name such as Yitang Zhang coming out, this can easily cause some confusions. For example, "Zhang's Theorem" really provides little information about the author "Zhang" since "Zhang" is too common. This terms can not only cause confusions (we may have "Zhang's Theorem" in every area of math in near future!) but also indicates little credit to the person who proved the theorem.
I suggest that when we cite a paper from a Chinese author or name a theorem, we should use full name such as "Yitang Zhang", instead of "Zhang" or "Y. Zhang"(even with "Y.", this is still too common). This should not be viewed as a discrimination nor a preference: this principle can be used whenever there is a risk of confusion (e.g. be more specific about "Smith", "Vinogradov").
PS: one motivation of my suggestion is this question: Which pair of mathematicians has the most joint papers?
One answer is "This is a frivolous item solely to demonstrate the pitfalls of running MathSciNet searches and working with large datasets:
Type "Wang and Zhang" in the author field and get a list of 2417 items. Li and Wang are close contenders with 2300 total. I wouldn't venture a guess how many collaborations that represents!"
I think now you can see why "Wang and Zhang" or "Li and Wang" collaborate so much.