Hi pals, Julia Weist from Collective Access shared some thoughts with me, and now I'm sharing them with you. Note: I'm not saying we should use CA for our catalog, just that Julia gives us some things to think about. Jenna ---------- Forwarded message ---------- We have some experience with consortium projects here at CA, and those function similarly to how a cooperative union catalog would. In those cases, we've had associations (comprised of several dozen museums) manage individual back-end systems where they separately enter their data. Those back-ends systems, though separate, are identical in terms of metadata and configuration. Each system sends data (through scheduled migrations--could be daily) into a single aggregator system. That pooled back-end fuels a single front-end. We've learned a few lessons with this, and I'll share those in case they are helpful! 1) We had to develop special tools which I'll call "metadata quality assessment rubrics" which use regular expression (/pattern matching) to flag data that was subpar based on the rules of a shared content standard. This helps your data look clean and consistent when it's being created from so many different cataloguers. 2) Every difference between systems adds complexity. In our case this wasn't a concern because we designed all the systems to be identical, but if you're working with various partners on various platforms it can quickly become a ton of crosswalking and custom sub-projects to get one system to talk to another. My advice would be to pick one shared platform and a common metadata schema. 3) The aggregation needs to be have built in rules to handle authority reconciliation. What happens if two people submit data about the same person on the same day? Who's is kept? Are they merged? If merged, will the data be redundant? Sometimes this can be entirely handled with a rule set, other times a human intelligence is required. CollectiveAccess could certainly work for this, because as I mentioned we've done it before so it wouldn't require reinventing the wheel. But configuration, build out, partner wrangling and support does take time, so we'd need funding if you wanted to go that route. In any case I'm happy to provide an opinion here and there if that helps, or to talk more seriously about using CA if you want. In either case sounds like a great project! Julia _______________________________________________ Zlunioncat mailing list Zlunioncat@lists.qzap.org http://lists.qzap.org/listinfo.cgi/zlunioncat-qzap.org