Digital library experiments

This is a list, in reverse-chronological order, of a number of digital library experiements -- demonstration applications illustrating various ways computer can be used to implement digital library collections and services against those collections:

• SRU2Aleph - This is a minimalist "hack" illustrating how the SRU software created by Index Data and implemented against Ex Libris's Aleph catalog can be used to create an additional interface to the library catalog. The program took about 60 minutes to write, and it looks like it. On the other hand, it illustrates a way our catalog content can be searched and presented through the use of standardized computing techniques, namely, SRU. The simple .html client returns raw XML. The .cgi client transforms that XML into something a bit more usable. (September 24, 2007)
  
  
• Reading List - This is a customizable list of open access journals -- yet another MyLibrary implementation. It's content, list of journal titles, comes from the Directory of Open Access Journals (DOAJ). After it is imported and automatically classified, a simple list of titles is presented. Users can then browse the lists and add (or delete) them to their "reading list".  Given such a list a person could refer to it on a regular basis and read the new articles listed there. While such a function is not necessarily one that needs to be solved, the application demonstrates how personalized services can be implemented with MyLibrary using free and open access content. (July 12, 2007)
  
  
• Another catalog - This is a MyLibrary demonstration -- a library catalog with the addition of a couple of bells and whistles. The catalog indexes 300,000 MARC records. The index is accessible via SRU. Search results sport cover art and reviews, when available. It also prototypes a "Get it for me!" function by allowing users to create accounts for themselves. Once items have been "checked out", the user can see what is on their "bookshelf". With a bit of statistical analysis, combined with authority lists, a person's "bookshelf" could be used to recommend addidtional titles, a specific librarian, or sets of bibliographic indexes for further use. (July 7, 2007)
  
  
• Article Index - This is a MyLibrary demonstration, and it serves two purposes: 1) to illustrate ways DOAJ content can be re-used and repurposed, and 2) to demonstrate how MyLibrary can be used to support digital library collections and services. DOAJ is an OAI-PMH data repository. It supports at least two different data sets: 1) metadata regarding journals, and 2) metadata regarding articles. We wrote a simple OAI-PMH harvesting program that pulls down the article-level metadata. We store the data in a MyLibrary database, and then provide searchable/browsable interfaces to the result. The browsable interface is rudimentary, including only publishers, sources (journal titles), and broad DOAJ-specific "subject headings". With more work a browsable author list could be created too. The entire content of the system is made searchable by indexing it with a Perl module called Kinosearch. This index is then accessible through a standard SRU (Search-Retrieve via URL) interface. This allows us to change our underlying indexer at will and not change the search interface. (July 1, 2007)
  
  
• Send It To Me! - Given an ISBN number this system will query a number of sources, return the item's bibliographic information, and allow the user to request that the item be physically delivered to them. If the item is not found, the system will offer the user the ability to find similar items and have them delivered instead. (May 1, 2007)
  
  
• Collection Manager - This tool allows collection managers -- bibliographers -- to browse OAI-accessible archives and identify possible collections to archive.  Behind the scenes a set of archives are indexed from the Directory of Open Access Repositories (OpenDOAR). The user can then enter a query, and a list of archives will be returned. The user can then drill down into the archive in search of content of interest to the Libraries. Once identified, the OAI URL's can be given to a harvester for downloading and indexing. (March 30, 2007)
  
  
• Open Content Index - Like Simple Search, below, the application allows the user to select an index and enter a query. Unlike Simple Search, the application does not redirect the user's browser to the remote index. Instead, Open Content Index searches the remote index(es) directly and returns the results. It does this through a search protocol called SRU. All of the content in this index is "open", meaning it is free and unencumbered from licensing agreements. (March 22, 2007)
  
  
• Simple Search - Select an index, enter a query, and be redirected to the remote index while having your query executed at the same time -- a simple search interface. Ideally the HTML form would be implemented as one line of Javascript allowing webmasters to include the form on any Web page across campus. (February 16, 2007)
  
  
• Catalog - This application searches the library catalog through the use of X-Services technology. As results are returned they are examined and statistically significant words are brought to the top of the results, and synonyms are presented for possible alternative searches. The entire results can be transformed into an RSS feed to provide a "What's new?" feature. (Feburary 5, 2007)
  
  
• Your Page - This is a MyLibrary demonstration. After importing patron information from the registrar, after importing lists of classes from the registrar, after creating sets of library-related information resources (indexes, catalogs, Internet sites, names of librarians, etc.), allow the patron to log in and see: a list of suggested resources based on their major/department, resources based on the classes they are taking, and a list of the books they have checked out. This is  the beginnings of a personalized (as opposed to customized) user interface. (January 15, 2007 -- Martin Luther King Day)
  
  
• Tomato Juice - This is a "mash-up". Enter a query and results will be returned from dictionaries, thesari, encyclopedias, library catalogs, image collections, and Internet indexes. Along the way the program saves queries and reformulates them as a "tag cloud" allowing people to see what others have used the system for and what is popular. This program illustrates a way to use Web 2.0 technologies to bring together information from distinct resources. (August 20, 2006)
  
  
• Fun with KinoSearch - This is an exploration into the use of an indexer called KinoSearch. Using this indexer I created a search interface to OAI-accessible content (content from the Directory of Open Access Journals and sets of images), as well as sets of MARC records. I then combined the three indexes into a master index to implement a metasearch. I think libraries should create more of their own indexes. (May 26, 2006)
  
  
• "Catholic Portal" - This is a part of the Catholic Research Resources Alliance. It includes pointers to unique and rare materials of interest to Catholic Scholars. The application works by reading EAD files, parsing them, and providing a searchable/browsable interface to the content. This too is a MyLibrary implementation. (September 18, 2005)
  
  
• OCKHAM Initiative - There are three applications associated with the OCKHAM Initiative -- a grant funed by the National Science Foundation. The first is called OCKHAM Alert (August 25, 2004). It regularly harvesst content from NSF OAI repositories. Saves the content to a database while deleting the data that is older than 30 days. Indexes the content. Provides an interface to the index, and allows the user to save the search as an RSS feed that can be read as desired. Ideally the content should change over time while the queries stay the same. The second application is called MyLibrary@OCKHAM (March 15, 2005). This the most extensive application. It harvests content from OAI archives funded by the NSF, automatically classifies the content, indexes the whole thing, and supports a searchable interface to the index supplemented with alternative spelling and synonym suggestions. It also support a "find more like this one" feature. This was one of the first SRU-accessible indexes. The third is called OCKHAM Spell (September 20, 2005). This simple Web Service was an outgrowth of the MyLibrary@OCKHAM. Given a word and a dictionary, it will return a list of alternative spellings in the form of an XML stream. Such an service can be embeded in other applications to implement things like Did You Mean? All of these projects were documented in an article, Exploiting "Light-weight" Protocols and Open Source Tools to Implement Digital Library Collections and Services by Xiaorong Xiang and Eric Lease Morgan.
  
  
• Gadgets - Listed here are a number if unsupported Firefox extentions that provide enhanced access to library collections and services through a Web browser. They include things like an OpenURL resolver and widgets to search specific indexes on campus as well as in the Libraries. (June 9, 2005)
  
  
• DOAJ Index - This is an attempt to: 1) mirror open access listed in the DOAJ, 2) index the content, and 3) provide an interface to the index. The process more or less worked, but administrative interface broke. This application illustrated a way to provide spell-check against searches, and it began to implement a user-defined customizable interface. Finally, YAMSI -- Yet Another Mr. Serials Implementation. (February 23, 2004)
  
  
• Microforms - This is a set of three indexes against sets of MARC records: 1), Eighteenth Century Literature, 2) Landmark, and 3) Newspapers. Each of these sets of records were acquired when the Libraries purchased the corresponding content. By indexing these records outside the library's catalog we could provide specialized search interfaces against them. Such interfaces, in conjunction with the interface provided by the catalog, could provide enhanced access to the content. (January 20, 2004)
  
  
• MARC Index - After locating a random set of MARC records, this application simply parsed them, fed them to an indexer, and provided an interface to the index. The interface supports browsing by author, title, and subject. The search interface supports fielded searching, truncation, sorting as well as relevance ranking. There is a preferences feature allowing the user to customize the appearence of the output. It also keeps a history of the user's queries so when the user returns later they can start up where they left off. This was an adventure in the creation of a locally developed online catalog. (January 4, 2004)

Date created: September 17, 2007
Date updated: September 28, 2007
URL: http://www.library.nd.edu/daiad/morgan/experiments