The Online Library of Liberty used to be published by hand, with every title hand-coded, organized and linked. It was a little bit like the printing presses of yore, where each character had to be set in place by hand. With the new system, this process is automated and organized to allow maximum efficiency, remove duplication of effort, and offer more value for the site’s visitors.
As a part of the Liberty Fund’s mission to encourage the study and discussion of liberty, each title in their collection is completely re-typeset, working from photographic copies of the often fragile original texts. Once printed, these new editions are then donated to universities and schools around the world. An archival XML file in the Text Encoding Initiative (TEI) format is created in the process, which captures both the structural nuance of the original pages and their content.
Adding a new title to the Library starts with the automatic conversion of the XML file to valid XHTML. Next, the document is split into paragraph-sized chunks and stored in a MySQL database for indexing and redisplay. Each paragraph is tagged with its place in the book, and can stand alone or be re-assembled into chapters, sections, or the entire document.
At every stage of the process, the interface is designed to avoid any duplication of effort and to remove opportunities for error. For example, any option that has already been entered into the system can be chosen from a list, rather than re-typed.
When all of the information about the title has been entered and proofread, one click publishes the title, adds it to all relevant indexes and lists, and makes it find-able through the search system. The new system makes it possible to add a title to the library in a few minutes, rather than a few days.
A powerful natural-language search engine is essential for a collection as wide and broad as the Online Library of Liberty. In addition to full-text searching within books—in multiple languages—the search engine also uses the bibliographic and biographical data to locate matches. Visitors can further customize the search to only match titles within a particular category.
The public site uses the Joomla open-source publishing toolkit to present the Library and to add and update editorial content.
Like the 9/10ths of an iceberg that you can’t see, the administrative interface is where most of the magic happens in this system. Liberty Fund staff use this secure Web interface to manage every aspect of the Library, from initial parsing of the XML files to the final touches needed by scholars around the world. Watch it in action!
Each title in the Library is associated with one or more people: authors, editors, translators and more. And each person or title can be a member of one or more groups, like Authors from the 18th Century, or Titles about the French Revolution. These associations allow the public site to dynamically create lists of related titles in any view. These associations also help the system to optimize the page content for search engines.