This document has been superseded by the Use Cases.

The GRN task is complex. The data collated here are of a wide variety which are as likely to muddle search results as to illuminate them. How can we guide the researcher without losing them in the complexity? That's the core of our challenge.

These use cases refer to illustrations in the use case PDF. This illustration and these use cases were built off brainstorming done on 090223 at MHS.

The Researcher Actor

Different varieties of researchers don't label themselves as they arrive at the GRN doorstep. As they make their way through results that GRN brings forward, they will distinguish themselves, but the really all start out as a "researcher" actor, so I don't distinguish here. This conflation of genealogist, librarian, student, teacher, family, and curator into one research is represented by the "*" in the PDF.

Use Case Drafts

A1. Developer seeks site map for own index.

Andrew wants to index the GRN data and provide a visual search interface. He uses the web services of GRN which provide a regularly updated site map that facilitates a regular and efficient crawl of GRN data.

A2. Developer integrates results into own site.

Allison loves the GRN search & retrieval, but needs to integrate a very particular search and its results into her legislative web site, which looks completely different from GRN. Instead of resorting to screen-scraping, she finds that GRN provides a tool for submitting searches and getting the results as JSON or XML packages.

B1. Developer seeks documentation of web service.

Betsy wants to allow people on her site to search Minnesota birth records. She's heard that GRN at MHS can provide links to such records as a service that she can integrate with her own web site. She finds documentation on how to integrate GRN results into her own page on the GRN site.

B2. Partner seeks documentation of contribution format.

Brad is responsible for preparing his historical society's data for ingest by the GRN. On the GRN site he finds documentation on how to prepare data for ingest, including links to documents which define the standards GRN staff use to move data between systems.

C. Partner sets up new contribution stream.

Brad has followed the instructions he found in B2 and is ready to contribute his institution's content to GRN. The final step is to announce the availability of the content by filling out a form on the GRN system. Carol receives the completed form and adds Brad's data to the harvest list and indicates that it should only be used on the test system for now. Once harvesting has begun, she notes that the status of this stream is "testing" and the GRN notifies Brad, who reviews the appearance of data in the test system. If all looks well, he approves the test data and the GRN alerts MHS staff that the stream is ready to go into production. Carol directs the stream to the production system.

D1. Admin adjusts maps type from partner stream.

David, an administrator of the GRN, gets a note from an MHS librarian pointing out that some maps from Metro Transit, a GRN partner, are not getting categorized as maps. David takes a look at the stream of data harvested from Metro Transit, which turns out to be a relatively simple web harvest. Inspecting the source of these pages reveals that the maps in question are always PDF format files with "rtmp" in the name. David sets up a new step in the Metro Transit harvesting workflow to add a "map" type to each item that meets this test. He reinitiates the Metro Transit workflow and soon these records are being recognized as maps on GRN.

D2. Partner removes abusive comment.

[Note: Is this just part of WOTR? If so, drop it.] Deb gets a note from the GRN that one of the comments on a Metro Transit collection item was flagged as "inappropriate" by another user. Since this hides the comment from public view, she knows that resolving the issue should be a high priority. She logs into the GRN with her partner id and goes to that record. The comment is there, and it looks pretty bad. Since Deb has used her partner id, the comment also has a "hide comment" button and a "confirm inappropriate flag" button. Both would result in the comment being hidden from public view permanently, but she uses the "confirm" button since that one also thanks the user who originally flagged the item.

E. Funder seeks monthly contribution and usage statistics.

Eileen is a staffer at the Bush Foundation and needs to prepare an update for the foundation board. She is looking for statistics about various grant project. She is pleased that she can find GRN usage stats on the site.

F. Researcher leaves a comment on a record.

[Note: Is this just part of WOTR? If so, drop it.]

G1. Legislator seeking the head of MHS exhibits.

Rep. Smitson wants to give an important constituent a couple tickets to what he has heard is a wonderful exhibit at MHS. His staff secretary, Glenn, has no idea who can help him, so he heads to the search box on the MHS page and types in "exhibit tickets". Since this search started on a general mnhs.org web page, the results highlight MHS information. Glenn sees a link to the exhibits tickets page in the results and clicks. Luckily he sees a phone number for group purchases on that page, close enough, he picks up the phone.

G2. Partner trying to find Bob's office.

Gail is coming to town for a meeting with Bob about their partnership with the GRN project. She's about to leave when she realized she does not have Bob's office room number. She heads to mnhs.org and types Bob's name into the search box. Since this search started on a general mnhs.org web page, the results highlight MHS information. Gail sees Bob's name in the results and clicking on it goes to his directory information at MHS. Unfortunately the room number is not there, so she calls Bob instead.

G3. Researcher (DGP) looking for Forestville.

Gaston wants to visit Forestville and all he can remember is that someone told him it was associated with MHS. He sees nothing about Forestville on the front page, but decides to try the search box. Since this search started on a general mnhs.org web page, the results highlight MHS information. Gaston sees a link to Forestville directions and clicks away.

H. Researcher wants to know more about "John Pool" in Minnesota.

Henry is a genealogist working for the Pool family. His research has led him to a branch of the family in Minnesota, so he comes to the MHS site seeking more information about the family. He starts by entering "John Pool" into the search box. Since this search started on a general mnhs.org web page, the results highlight MHS information. But Henry notices that one of the smaller columns show birth and death records from Minnesota. He clicks the "more of these" link under the birth records. The search results rearrange themselves, making the birth records more prominent and providing an opportunity to enter first and last name separately. He takes advantage of that, and ends up with a very helpful summary of people data from GRN.

I. Researcher looking for pictures for history day.

Ingrid is looking for images she can use as part of her history day essay on the St. Anthony Falls. She types "St. Anthony Falls" into the search box on the MHS home page and ends up with a page that has all sorts of results on it. Right at the top is some link to an exhibit years ago at the History Center, but what catches her eye is a small grid of pictures further down the page. Since she's looking for images, she clicks on this grid and ends up with a page that makes a whole set of images more prominent in the results. She starts sorting through them for ones that would work for her paper. Ingrid has to hurry since it is due tomorrow!

J. Researcher saves a search.

Jeremy has spent some time constructing a nice search of GRN for his topic. He's used some limiters and some boolean terms, and he's afraid that he'll forget the combo in the coming weeks. He'd like to check back later to see what turns up. He clicks the "save search" button and gives the system his email address. The system sends Jeremy an email message with a link he can use to re-execute the search at a later time.

Implications for Functionality

The types of information which may be revealed by a GRN search are so various that we will need to determine some dimensions upon which to sort them out and present them in sets. The current GRN implementation makes one attempt at this, largely dividing the material based on the source from which it originates. Presented as one long list of sources with short lists of results in each one, this gets to be a bit debilitating for a user. They have to think too much about the dimensions.

It seems that a significant portion of our challenge in GRN2 is to think carefully about dimensions of information that might work for the user. In the brainstorming the dimensions of "citation" vs. "object" kept coming up. Another might be formats: image, audio, text, maps, video and so forth.

These dimensions also need to be presented to users. The current list strategy seems to lose people. I wonder if an approach like http://kosmix.com might serve us better. Easier said than done, obviously. Though at least we are dealing with a _less_ complex environment than kosmix!

The workflow management of gathering streams of data into a single index while adding certain categorization to these data based on their source or attributes of their contents is non-trivial, to say the least.

The positive nature of the response to the current peoplefinder also has me a bit spooked. How well will we be able to keep this success intact away from IDOL. I know nothing of IDOL or how it has facilitated these results, but key aspects of this functionality should be specified for GRN2 as well.

Comments

Please add your thoughts!

Marj / 03 March 2009 / 09:27

From what I recall the web service actor is to enable web crawlers and other applications access to the data so it can be used in their own systems or applications. For web crawlers it would create a site map to the data. For other apps it would return results that could be used for mash-ups or in other ways by other apps, which you do have in the B1 Developer scenario.

I’m not sure that the G1. Legislator or G2. Partner scenarios are any different as described than a researcher. My guess is what makes them unique is that they would want access to the usage statsitics like E. Funder. The details of the G1 and G2 cases however maybe useful as the “dreaded general public” searches. It may also be useful to have a generic person search for information about an exhibit or event at Mill City or the hours and directions to Forestville.

The kosmix site is interesting. I’m going to use it instead of Google for a few days to get a better feel for how it works. Thanks for point it out.

efc / 03 March 2009 / 14:15

Great ideas, Marj. I've tried to incorporate in A1, A2, and G3 for discussion tomorrow. I also removed this statement about web services, though I'd still welcome other thoughts on the topic:

What are some examples of web services that folks might use? I'd thought of RSS during the brainstorming, but RSS makes little sense on a system which will re-harvest and re-index regularly (no way to tell old from new content). So what other examples do we have? Maybe other state agencies want to put a "peoplefinder" search up on their sites? Maybe simply integrating search boxes to GRN? Maybe saved searches that go to deep results? Ideas?