But no really, be practical, HOW will we do all that?

This week I also was unsure what to blog about; I considered my most frequently scribbled marginalia. The question is quite simply, how? It may be a mundane question, but I am a person who can't plan without understanding it.

Here's one example from this week: "Data at both the individual and firm level, as well as national and international levels, must be integrated. Complex qualitative and quantitative data from a variety of modes must be combined." (p. 20 of http://ucdata.berkeley.edu/pubs/CyberInfrastructure_FINAL.pdf)

While I am aware that technology changes quickly, I thought it might help to find some examples of how current digital library projects are structured. I began to hunt for articles about social science databases. Megan's comment about persistent identifiers was timely; 404 not found and I got to know each other quite well.

Finally, I found an article about the Exploratorium, an online science museum. http://saturn.exploratorium.edu/partner/nsdl/pubs.html
The article, From Playful Exhibits to LOM: Lessons from Building an Exploratorium Digital Library, can be found here. The authors describe the metadata standards adopted to create the interactive Exploratorium.

Specifically, they designed their metadata for use with "Learning Objects." They knew in advance what their mission was, to provide educational material, so they knew the metadata should be structured to support that. The standard they chose (IEEE LOM) also maps to Dublin Core, so I think it is extensible in the way that Berman and Brady were suggesting from my earlier quote.

They assessed what they had to put in the collection before starting, and then used an assets management solution called Canto, which supports scalability and APIs.

Should they have tried to combine more projects and possibly be more extensible? Do they have enough tools, enough scope, or enough added information to be classified as a curation project? I'm not sure. I know I found their website much easier to understand than some others. An image of one view of their structure and metadata is below.

The question of "how?" deserves more consideration, but I believe this site is a good example of knowing what you want in the collection, setting goals and standards, and adapting when necessary. The authors acknowledge that their use of URLs as stable locations may need to change, but currently they do not have a better option.

(I did not receive a single 404 error while exploring their site.)