The Web is looking for intelligence, however the traditional knowledge representation methods are inadequate to facilitate human-machine interaction since they mainly focus on the results of the interaction, i.e. the knowledge in the "object" form. Based the analysis on machine languages and natural languages, we argued that a subject-oriented knowledge formalization supported by controlled natural language is an important way to bring human intelligence to the Web. Thus human-machine interaction should shift into the "What You Think Is What You Get (WYTIWYG)" paradigm. This user friendly method is useful when a large number of personal experiences are needed for sharing and reuse.