The expansion of personal information archives and the emerging field of Personal Lifelogs (PLs) are creating new challenges for information retrieval (IR). While studies have demonstrated the difficulties of IR for these massive data collection [1], we should also think about how we can opportunities and benefits from integrating these data sources as a component of “digital memories” , considering their rich connections with the users‟ memory. We observed that most existing approaches to personal archive IR are mostly technology-driven. Although in recent years studies in Personal Information management (PIM) have claimed to make use of the human memory features, and many works have been reported as investigating well-remembered features of computer files (documents, email, photos). Yet, these explorations are usually confined to the attributes or feature that current computer file systems or technology have provided. I believe that there are important and potentially useful data attributes that these studies have ignored. In addition, current personal search interfaces provide searching options based on what is available in the system, e.g. require users to fill in the calendar date, regardless of the fact that people actually don‟t often encode „time‟ in such a way. My PhD project aims to explore what users actually tend to recall in different personal achieve information seeking tasks, how to present searching options to cater for the right type or format of information that users can recall, and how to exploit this information in an IR system for personal lifelog archives. In this paper, I discuss the limits and advantages of some related work, and present my current and proposed study, with an outlook of an interface that I plan to develop to explore my proposals.