Montana State Library has about 4000 objects (state publications) to upload to the Internet Archive and Open Library. Would like to automate as much of this as possible. For each object, I need to produce to an upload folder named with unique identifier. In this folder, among other things, I need to put the upload object and a metadata xml file which I intend to crosswalk from a marc record (a .mrc file). We are SirsiDynix shop (running on top of Oracle) and otherwise have a windows environment (but would like to hear about whatever process works regardless of platform). I am attracted to the z3950 interface for pulling marc record data which I then propose to crosswalk to the metadata xml file required by the Internet Archive. Getting the .mrc file can be done manually using a z3950 client. However, I want to address the z3950 client in a script loop and pull and process .mrc data for each object. Can this be done? It appears yaz can be controlled from the command line and that I might be able to loop it using the shell? Not sure however if this can be done from a dos shell. If not, other thoughts are in order about what might be the best way to do this? Thanks ...