I think the post title about says it all. March is the birthday month for the University of Florida Digital Collections (UFDC) and this year UFDC turns 3!
Where most three year olds aren’t yet learning to read, UFDC is already onto using Lucene for indexing (it’s like reading, so this metaphor sort of works) and UFDC already has well over a million pages for each year. Those millions of pages weren’t added evenly throughout UFDC’s early year and a half or so, but UFDC is now growing and thriving.
As we look forward to the next year and the next series of millions of pages, many new plans are in development. Some are more exciting like the Caribbean Newspaper Digital Library proposal, and some are less glamourous but still rather exciting like the move to put UF’s internal FilmLog database online so that others can see just how many reels of microfilm of what are available. I know microfilm doesn’t sound exciting, but many materials are no longer available in paper (at least in Florida, thanks to the weather) so microfilm may be the only copy. Plus, we’ve already digitized and are processing bunches of reels of that microfilm so it’s good for people to know what else might be coming and what else they might want to know more about.
In addition to more content, UFDC’s third birthday is a time to look forward to many new technological improvements and many more ways to share and explore the millions of pages of content, audio files, video, and images with others.
Happy Birthday UFDC!
Mark sent me a note on some of the UFDC/DLC accomplishments this year, so I thought I’d add them here in celebration of UFDC’s birthday!
UFDC Web Application
The UFDC web application was optimized through a four month process of examining every bit of generated HTML, every line of C#, and performing memory optimizations. As a result of this work, the HTML output is significantly more compact and the number and size of images served was reduced. Overall, this increases the speed of UFDC for users with lower bandwidth. In addition, client-side java scripts were utilized to reduce the number of round trips to the web server, increasing performance and speed again for the users. A slow persistent memory leak was found and corrected during this process as well, which had been causing occasional service outages. The user interface also underwent format changes in response to usability tests performed against the interface.
The architecture of UFDC was examined with major changes implemented on the item display portion. Previously, the item image information was queried from the underlying Greenstone layer. With the new changes, Greenstone is only utilized for searching against the metadata and text. Once an item is requested for display, the METS file is read from the web server and the item is rendered from that information directly. This, in addition to optimizing the METS reader to increase speed four-fold, allows for much quicker item retrieval and more consistent display throughout. In addition, having the full metadata at hand allows the citation to be displayed in MARC format, which allows librarians to review the metadata in their native tongue.
Several other changes were implemented within UFDC. These include the addition of a robust statistical reporting module to report on usage of UFDC since its inception. Also, new item viewers were added, such as the Flash viewer. The underlying Greenstone layer was upgraded to version 2.81 and indexing moved from MGPP to Lucene.
UFDC Builder
The UFDC Builder was created, merging the tasks of the previously separate UFDC Preloader and UFDC Loader. This merging reduced code redundancies between two separate projects. In addition, the Builder was created to be running fairly continuously. It monitors the FTP boxes every ten minutes. Metadata updates are processed within 10 minutes of being submitted. With the change to the item displayer mentioned above, this means that metadata updates can often be reflected online within a matter of minutes, rather than a full day later as before. In addition, new items are continuously being added to collections and to the browses, again increasing the speed with which items appear online.
The UFDC Builder also facilitates increased access to our resources. As each item is added to UFDC (or altered through a metadata update) static pages are created by using the same library that the UFDC Web application uses. These static pages appear exactly like the corresponding dynamically-created UFDC pages but are made available for indexing by search engines. In addition, RSS feeds are generated at the collection and institution level within minutes of new items being loaded. And finally, MarcXML files are generated to increase interoperability with other library systems, such as the state-wide library catalog. All of these changes have resulted in a large increase in usage of the UFDC web application itself and increased visibility for the University of Florida Libraries.
dLOC Toolkit
Through examination and on-hand usability studies, it became clear that the original dLOC Toolkit was too difficult for many users. This toolkit was discarded and a new version was created. The new version (Version 2) is significantly simpler to use and guides the user through the digitization process in a very straightforward manner. This has increased contributions to the Digital Library of the Caribbean and decreased the amount of time needed for training new partners. This new version has piqued interest in many disparate groups. Again, this work has increased the visibility and perceived technical capabilities of the University of Florida Libraries and aided researchers and librarians in their work.
Happy Birthday, UFDC! You grow faster than my puppy.
–Missy