Monday, December 15, 2008

Service Infrastructure Hack

Here's a diagram I drew a couple of years ago, when I was first trying to figure out the core services that make up our technical infrastructure. If I did it again today, I'd do it differently, but in the interest of discussion, I'm posting it as-is.

Decoding some of it: "ILS" is Integrated Library System; "CISR" is Comprehensive Index/Search/Retrieval; "IR" is Institutional Repository. If I re-did it, the CISR and IR functionalities would get blended together and then un-mixed into something like "Digital Collections" and "Digital Content Management".

Blue arrows represent all of the harvesting that turns the CISR into an FRD ("Federated Resource Discovery") service. Red arrows represent peering relationships with sister institutions that in turn elevate an FRD into a GFRD ("Global FRD").

Anyway ... worth what you paid.


Edie Dixon said...

I think I don't really know, anymore, what we mean by "federated"; whether we apply that word to "search" or to "discovery" (or government for that matter but that's a whole 'nother topic). Are we trying to present a box that somehow takes the input and just plugs that input into everything we can think of that a person may want to know about? Or do we want to be a little smarter than that? Take what's input, make some guesses about intent, find a couple of paths to search that match that intent, and then give back results? Is that "federated"?

When we present a search box to people and say "put some words here and see what you get" we may better serve folks by assuming it really *is* a subject, and then call a service that does subject searching right away.

I guess I think that we should pursue what people think they are typing when presented with a search box -- I'm not sure we know, really.

Mike said...

I think all I mean when I say "federated" is that each of "us" (CIC schools, UW campuses, whatever) can harvest things from all the rest of "us". "Global" just means that I can do that harvest using similar means/protocols/etc. across the different kinds of information that the library curates: bibliographic, digitized content, citations, the whole chile relleno. I think your questions are actually more important: once we have an infrastructure that lets us do that, lets us move information smoothly and cleanly between the different buckets that collect it, the choice of how we want to guide people through it, the kinds of interfaces that we want to present to help them manage all of that data -- that's going to be where the real artistry lies.

Peter said...

I'm not sure what the distinction would be between "digital collections" and "digital content management" - content management seems like something you do with collections. The existing distinction seems right to me - I'll post my picture of part of this same landscape.