一句话：回顾约束满足问题（CSP）复杂度二分猜想的提出背景，说明为何要在广泛问题类上研究 P/NP 二分现象。
该文讲述约束满足问题（CSP）复杂度二分猜想的背景故事。作者首先指出，长期观察发现“自然”计算问题常完全于 NL、P、NP、PSPACE 等自然复杂度类；Ladner 在 1975 年证明若 P≠NP，则二者之间存在无穷多复杂度类，但已知中间问题多依赖对角化构造且十分人为。Feder 与 Vardi 的开创性工作使这一现象被称为（P 与 NP 的）复杂度二分，Valiant 在计数问题中也有相关发现。由于“自然”问题难以精确定义，一个研究方向是对足够广泛的问题类证明二分结果，而 CSP 正是这样的问题类。文章据此勾勒 CSP 二分猜想的研究动机与历史脉络。
It has been observed long time ago that `natural' computational problems tend to be complete in `natural' complexity classes such as NL, P, NP, or PSPACE. Although Ladner in 1975 proved that if P ≠ NP then there are infinitely many complexity classes between them, all the examples of such intermediate problems are based on diagonalization constructions and are very artificial. Since the seminal work by Feder and Vardi [8] this phenomenon is known as complexity dichotomy (for P and NP), see also Valiant's work [14] in the context of counting problems. Concerted efforts have been made to make this observation more precise, and since the concept of a `natural' problem is somewhat ambiguous, a possible research direction is to pursue dichotomy results for wide classes of problems. The Constraint Satisfaction problem (CSP) is one of such classes.