This paper discusses the indexing, collection, and maintenance processes, as well as the retrieval interface, to public digital libraries. The paper begins by briefly surveying both Internet technical report collections and World Wide Web search engines. It goes on to describe a prototype digital library, the New Zealand Digital Library for Computer Science Research. The indexing and retrieval mechanism is described, including the full-text index. The types of search provided, the indexes that are needed to support them, and the query interface through which they are made accessible are summarized. The process of building the collection, and the crucial issue of what information is stored centrally and what can remain distributed, are then discussed. The final section summarizes how the information is gathered, considers the factors that limit the size of the library, and reviews future directions for networked digital libraries. Two tables contain information on technical report indexes on the Internet and the current size of the New Zealand Digital Library for Computer Science Research. Four figures illustrate the interface and query page to the digital library, conversion from PostScript (most text missing), the prologue used to extract text from PostScript documents, and the architecture of the digital library. Contains 12 references. (DLS)