Refactoring affordances in corporate wikis: a case for the use of mind maps

The organisation of corporate wikis tends to deteriorate as time goes by. Rearranging categories, structuring articles and even moving sections among articles are cumbersome tasks in current wiki engines. This discourages the layman. But, it is the layman who writes the articles, knows the wiki content and detects refactoring opportunities. Our goal is to improve the refactoring affordances of current wiki engines by providing an alternative front-end tuned to refactoring. This is achieved by (1) surfacing the structure of the wiki corpus as a mind map, and (2) conducting refactoring as mind map reshaping. To this end, we introduce WikiWhirl, a domain-specific language for wiki refactoring. WikiWhirl is supported as an extension of FreeMind, a popular mind mapping tool. In this way, refactoring operations are intuitively conducted as actions upon mind map nodes. In a refactoring session a user imports the wiki structure as a FreeMind map; next, conducts the refactoring operations on the map, and finally, the effects are saved in the wiki database. The operational semantics of the WikiWhirl operations follow refactoring good practices (e.g., authorship preservation). Results from a controlled experiment suggest that WikiWhirl outperforms MediaWiki in three main affordance enablers: understandability, productivity and fulfillment of refactoring good practices.     

并行环境下Java哈希机制的对比及重构

针对并行环境下出现的非线程安全问题,以哈希机制的Hashtable和ConcurrentHashMap的线程安全问题为切入点,设计了可改变线程总数和读写线程数的测试程序,分析了线程安全的ConcurrentHashMap和Hashtable的性能差异。根据性能差异,设计了从Hashtable到ConcurrentHashMap的程序自动重构插件,进而在Hashtable封装数据的情况下,通过对比传统多线程处理方式和Java8中的并行流处理方式的性能差异,分析了流处理方式相对普通多线程处理方式的高效性。对比结果显示,ConcurrentHashMap在并行程序中的性能要优于Hashtable。研究结果为多线程程序开发提出了解决非线程安全提供了较为可行的新思路,对于并行编程具有借鉴意义。