On 15/05/2012 16:50, Jonathan Rochkind wrote:
I certainly agree that the way our data is currently recorded and maintained in MARC is not suitable for contemporary desired uses, as I've suggested many times before on this list and others and tried to explain why; it's got little to do with rdbms though.
Although MARC needs to change, and has needed it for a very long time, I don't see how changing the format would improve the subject headings. The semantics are there already, so searching would remain the same. It is the display of the multiple search result which has disintegrated. I think there are lots of ways that the displays could be improved for the public--primarily by making them more flexible and could be experimented with now--but even then, there will need to be a major push from public services to get the public to use and understand what the subject searches are. All of it has been effectively forgotten by the public.
For a whole lot of reasons, library subject searches will always be substantively different from what what people retrieve from a full-text search result and while librarians can understand this, it is a lot harder for the public.