The ultimate goal of information seeking is to provide people fast, accurate, direct, informative knowledge through seeking. Inspired by the bright future of this research area, some perspectives are worthy to mention.
	since the Web is so popular and knowledge in the Web is so abundant, there must be a way for Natural Language Process research to make full use of the Web and enhance system performance. However there is not much research done on this intersection.
	Currently, majority information on the Web is text data. Speech, or even nonverbal signals (posture, expression) should be the inputs for an information seeking system, including the Web.
	Other than mental model, which is really hard to measure, we could measure some physical indices (heart beat rate, impulse rate, temperature, brain wave, etc) about a particular user. It will be very helpful to determine a user’s mood and mental model.
	People begin to use the Internet search is because they are unsure about some thing. Therefore ambiguous user requests or inputs are very common. What if we could offer a system to conduct the interview process like in the traditional library that the librarian answers a few of questions to guide the user to clarify his own request? Such system will definitely understand the user’s requirement better and hence provide better services.