Take a database conference paper and search for "in the real world" in it; chances are high you will find it. Of course what is real depends on one's perspective: for a pure theory paper it could be what one saw in a systems paper, for a systems paper it could be an issue that implementors of DBMSs had to deal with, and for the latter it may be what the customers need. But to sharpen our research tools, it helps tremendously to understand that the real "real world" is, and adjust our (sometimes very elaborate) techniques to address problems that actually occur. A nice example of this is analyzing the computational complexity of database queries. Database theory has developed an arsenal of tools for this. We know that for many classes of queries, the complexity is roughly kDkO(kQk) for a database D and a query Q, where k k means size. Thus, much research went into the detailed analysis of the structure of queries that removes kQk from the exponent and replaces it by a small fixed constant. We have a very good theoretical understanding of such classes, but we know much less about their relationship with queries that real-world users write.

中文翻译：

---

研究亮点

拿一篇数据库会议论文，在里面搜索“in the real world”；你会发现它的机会很高。当然，什么是真实的取决于一个人的观点：对于纯理论论文，它可能是人们在系统论文中看到的，对于系统论文，它可能是 DBMS 的实现者必须处理的问题，而对于后者，它可能成为客户所需要的。但是为了提高我们的研究工具，了解真正的“真实世界”是非常有帮助的，并调整我们的（有时非常复杂的）技术来解决实际发生的问题。一个很好的例子是分析数据库查询的计算复杂性。数据库理论为此开发了一系列工具。我们知道，对于很多类的查询，数据库 D 和查询 Q 的复杂度大约为 kDkO(kQk)，其中 kk 表示大小。因此，许多研究都对查询结构进行了详细分析，即从指数中删除 kQk 并将其替换为一个小的固定常数。我们对这些类有很好的理论理解，但我们对它们与现实世界用户编写的查询的关系知之甚少。