In this paper, we investigate the evaluation of the logs in software source code. Logs usually play the critical role in detecting, tracing and removing bugs of the software in the software development process. More importantly, after the delivery to the customers, software developers have to rely on the log records to locate bugs other than reproduce the software defects in the customer's environment. Therefore, there are strong quality and efficiency demands for good logs in the software source code. We model the source code as a hierarchy of graph models by presenting the relationships of different components of the source code, namely function call stacks, control flows and data flows. Then the fundamental metrics can be generated by the static analysis of these models and the metrics will be further processed to form the measurements of the log statements in the following dimensions: correctness, efficiency, maintainability, extensibility and conformity. We use a few open source projects to demonstrate the effectiveness of the proposed approach and discuss the directions of the future research.