Posts by Kat Hagedorn

Relevance is a complex concept which reflects aspects of a query, a document, and the user as well as contextual factors. Relevance involves many factors such as the user's preferences, task, stage in their information-seeking, domain knowledge, intent, and the context of a particular search. Tom Burton-West, one of the HathiTrust developers, has been working on practical relevance ranking for all the volumes in HathiTrust for a number of years.

(by Kat Hagedorn, Christina Powell, Lance Stuchell and John Weise) The one constant in digital preservation over the past couple of decades has been change. Digitization standards have changed as equipment has improved and become more affordable, formats have come and gone, and tools have been developed to help with automated format creation and validation. The progress made on this front has been great, but how do we reconcile older content with current digitization and preservation standards?

The Digital Library Production Service (DLPS) has created and hosted a very large number of digital collections over the last 10+ years. We have been working for many years to integrate those collections into MLibrary services, and we are now ready to present the next link in this chain-- a more easily navigable and more fully featured list of these collections.

The Usability Group & its Usability Task Force conducted a series of evaluations of the Library Gateway during the Fall 2009 and Winter 2010 semesters. We used a number of different methods, some new to us, to conduct our evaluations, including Participatory Design, Card Sorting and Guerrilla Tests.

We have recently made a number of significant updates to the HathiTrust Digital Library.

We have fixed a mistake with the UMProvider (OAI provider) that caused there to be more Dublin Core format records than MARC format records.

We have been making improvements to our OAI provider (UMProvider). We host the metadata for HathiTrust public domain texts through the provider, as well as all the metadata for text and image collections in the UM Digital Library.

Our first improvement was to make it faster to harvest. Our provider uses mySQL tables to store, sort and provide access to the metadata. Our method for sorting the data was one of the causes for the slowness of the harvesting.

Our second improvement comes from our investigation into the increasing number of deleted HathiTrust records that were showing up in the provider, and a discrepancy between the number of records in the provider and the number of records in our HathiTrust databases. We have not fully determined the cause of this, but we have been able to restore over 30,000 HathiTrust records that were marked as deleted in the provider.

Consequently, we recommend you harvest the provider from scratch, whether the entire metadata set or a particular set. It will be quick, and you'll get those missed records. We will keep you posted on further improvements.

(The UMProvider can be accessed via http://quod.lib.umich.edu/cgi/o/oai/oai?verb=ListRecords&metadataPrefix=oai_dc. There is useful information about the HathiTrust records in the provider at http://www.hathitrust.org/data.)

We've made some changes to the University of Michigan OAI data provider. The data provider now reflects the fact that we are providing records from the HathiTrust Digital Library (http://www.hathitrust.org/), formerly called MBooks.

Our recent article in D-Lib Magazine is a follow-up to the McCown et al. article in IEEE Internet Computing two years ago, in which the researchers investigated the percentage of URLs from OAI records in Google, Yahoo and MSN search indexes. We were interested in whether Google in particular had increased the number of OAI-based resources in its search index.

Over the past year we've been developing a new collection building tool to be used in conjunction with the MBooks "page-turning" application already available. This tool will allow users to create their own collections of MBooks items and view public collections created by others. Users will also be able to do full text searching across all items within a collection.