About a year ago, Special Collections took on responsibility for indexing the local newspaper, which put a dedicated indexer in the department. Working with him, I decided that, as part of our convrsion from beta to v1.0, we should add controlled access terms at both the collection level and series level of our inventories. We're now in the process of clarifying our rationales so we can create policies and procedures for that process, and have hit a real swamp. So once again I am turning to the list looking for advice and examples from others as to how they handle the issues we're facing. We've established a couple general rules: * always use vocabulary from standard lists: AAT for <genreform> and various LC lists for others, creating local lists only for personal names and corporate names not part of the LC system. * always add terms for implicit concepts not explicitly expressed * add terms for as many perspectives as the collection requires. Guided by the list of MARC 6xx fields, we're routinely doing an intellectual scan for personal names, corporate names, conference names, geographical names, topical issues, genre of document, governmental functions, occupational titles. * add a controlled term only once at the collection level and repeat it only once at each series level as appropriate. We don't have the staff to regularize every entry of every name, for instance. But even this modest approach has created a problem in our political collections. Political papers, especially, contain information on myriad topics and myriad corporate names, in the form of committees, sub-committees, as well as geographical names. Further, the various files are typically well labeled, and the full-text searching features of our DynaText/DynaWeb software makes them highly identifiable in that mode, so one is tempted to leave them untagged. On the other hand, tagging has great power to provide order to the chaotic world of full-text searching, and it feels like we're shirking our professional obligation not to tag each name and topic at least once per collection/series level. On the other hand, we have only one 50% person applying 20% of that time to this task. A year ago, before the indexer joined our staff, we regretted that we couldn't staff any effort to apply controlled access terms, but had decided to live with that. Now that we have an indexer, what can we reasonably expect? We've talked ourselves around and around this problem, alternating between the highly professional and the highly practical approach, and we have real reservations about committing to either extreme. We're looking for a middle ground. What have others done about this problem? Elizabeth H. Dow, Ph.D. [log in to unmask] Special Collections -- Bailey/Howe Library University of Vermont "Dare to be stupid."