Patchwork Prototyping a Collections Dashboard
- May 27th, 2009
- Richard Urban
In days of yore, the researcher had a limited set of tools at their disposal to get a broad sweeping view of what a research collection consisted of. There might be a well-crafted NUCMC entry, a quick glance at a finding aid, a printed catalogue, or a chat with an archivist or librarian. Sometimes these pieces of information might tell you how much of a collection might meet your research needs (and correspondingly how many days you should plan to spend working with a collection).
Unfortunately many of our digital collections still rely on modes of presentation and description that are based on analog interfaces to collections. With increasingly large repositories (gathered into even larger aggregations) it is often hard for the researcher to know just how deep a particular rabbit hole goes. Improved search capabilities help solve part of this problem, but they can often impede serendipitous discoveries and unexpected juxtapositions of materials.
As part of our work to update the IMLS Digital Collections and Content project’s Opening History site, we are exploring ways that we can make the contours of a collection more explicit, develop modes of browsing that facilitate discovery, and provide researchers a sense of what’s available at different levels.
I’m looking forward to THATCamp because this looks like a great group of people to brainstorm with. Thus far, we’ve been using a paticipatory design technique known as “patchwork prototyping.” By the time of THATCamp we’ll have a few pieces of prototype together for review. If others are interested, I would be willing & able to lead a session that explores the general problem space using Opening History and any other collections that participants suggest.
May 28th, 2009 at 1:33 am
This sounds like a fascinating topic. We could have a pretty interesting discussion of how information gets encoded, represented, and recognized in finding aids for archival and manuscript collections. I think you raise an interesting question of how aggregation impacts search techniques. Your exploration of scale is on the right track, in my estimation. What might be missing from your model, though, is an exploration of the ontologies necessary for making the most sense of finding aids. In fact, I am of the opinion that in this era of Google-influenced keyword searching, researchers have actually gotten better at finding sources that exactly match their topic and worse at uncovering sources that are semantically related to their topic. We could talk about whether this is the case and what the remedy might be.
May 28th, 2009 at 8:46 am
Very interesting….it might be related to the transformation from info being hard to dig up to info being all too easy to dig up. When it was hard to dig up, one had to cast a wide net and focus in from there–a process that necessarily revealed those contours. With info being easy to dig up, there’s more need for filtering right off the bat, which immediately obscures the contours?
June 3rd, 2009 at 5:40 pm
@Sterling: I think you’re right about a need to explore the ontologies (both explicit and implicit) that are currently used to describe collections (and I’m sure Mark M. might have a few things to say about EAD), but at the same time that might be too distracting for THATCamp. At least personally I’d me more interested in exploring some grounded hunches that contours matter.
Do you think that Google’s precision is de-skilling researchers or just obscuring things which are relevant (but do not have a high page-rank?)
June 3rd, 2009 at 10:46 pm
@Patrick: I think there are also new possibilities that open up when we start having large quantities of item-level metadata around for collections which were never described at that level of granularity before. When collections are described at higher levels, there is a certain amount of purposeful wandering that’s necessary to get the lay of the land.
I also think there is something very human-scale about working with certain kinds of physical collections (even for large archives, we still talk about “linear feet” of materials – a very human-scale measurement). But digital collections are less tangible (measurements often rely on comparison to physical collections – how many times is the Library of Congress invoked as a yardstick for the capacity of some new digital medium?). I see a collection dashboard as a way of quickly and easily grasping the the quantifiable features of a collection on a more human scale.
June 4th, 2009 at 12:01 pm
I agree that ontologies can be time consuming. Maybe what would work best is an experiment in crowdsourcing, where participants in a session have five to ten minutes to work out an ontology for the digital humanities. This could at least raise awareness about the value and challenge of creating ontologies.
You can find some of my thoughts on how Google has affected research here:
cliomachine.org/wordpress/?p=65
I think young researchers who grew up on Google have lost some of the skills acquired by previous generations of researchers. Our concepts and measurements of relevant search results has also been watered down as Google’s influence has increased, in my opinion.
May 4th, 2010 at 2:51 am
[…] thatcamp.org/2009/patchwork-prototyping-a-collections-dashboard/ […]