This paper presents a good, realistic overview of XML, its potentials and advantages, and the research problems it presents to the database community. The last half of the paper is dedicated entirely to Widom?s own research project at Stanford University called Lore (a semistructured DBMS that was eventually migrated to an XML-oriented data model). Widom maintains an objective viewpoint throughout her paper - even when discussing Lore, she seems to speak without bias, describing both the strengths and current weaknesses of the Lore project. As Widom states, Lore does appear to be a good testbed for research in the XML area, and the research directions that she suggests seem very interesting. Although they are mostly described in the context of Lore, a lot of the concerns that are mentioned are certainly applicable to any implementation of XML. This paper touches upon a wide range of topics - storage and indexing, DTD?s, query languages, and general questions pertaining !
to the future of XML. This is a great paper for those who wish to get a broad idea of what research opportunities currently exist in the relatively new area of XML.