The idea of documents is important to this company. There is a great deal of interest in doing research in documents but given the very broad nature of the topic, it has been difficult to focus in a directed fashion. I am proposing A Research Paper server a vehicle to help provide this focus. All researchers at PARC use research papers to keep up with their fields. Informally published and distributed documents are much more important than books. Yet there are alot of problems with this medium. Research papers are distributed in a very informal manner. They are often copied many times. The version that a researcher has on his desk is not neccessarily the most recent. Many times the researcher aren't even aware that a paper is out. A machine that would keep copies of these papers would serve a useful purpose for PARC and would pose an interesting research problem that is important to the core of Xerox strategy. This server (a server could be a single machine or a distributed set of machines) would need to be able to scan in papers, store them, be able to retrieve and display them on a user's screen, and be able to print them. The document could be stored as imaginal or textual data, or both. From this starting point, several research issues could be addressed: The question of document representation becomes very important. There is a proposed OSI standard, interpress, various ad hoc proprietary (Tioga, TEdit) formats. The format question would need to be addressed. This could be a mchanism to understand the real issues. Physical Storage: To begin with, existing ÊB˜J˜ŠJ˜J˜ÄJ˜J˜¶J˜J˜žJ˜FJ˜J˜‹J˜J˜+—…—H