Recently, the Chronicling America Web site has been upgraded. Much of the interface is identical to the previous Web site version with only minor adjustments. However, some major changes in the underlying architecture will make the site more user-friendly overall, allow NDNP to expand capacity and improve performance substantially.
The improvements to the site from the user's point of view are primarily:
- All "Search Pages" results are now returned, by default, as graphic page images with hit highlights. Previously, the default was to the textual "List View" of results (title, date, page number). The "List View" is still available as a link on the Search Results page.
- All Web pages are now "bookmarkable" and persistent (see below for additional explanation).
- Improvements to the "Search Pages" date range calendar widget as well as other tweaks and adjustments as we rebuilt the application interface.
In addition to these small interface changes, recent changes to Chronicling America include not only functionality targeted at human users, but also programmatic interfaces designed for use by software agents. The three main changes are: use of bookmarkable URIs on all pages, addition of OpenSearch and Atom APIs for search queries, and and linked-data views available as RDF/XML (Resource Description Framework/Extended Markup Language).
While these changes are largely invisible to the human user of the system, they are visible in the site's HTML code. Browsers that are aware of syndication feeds will pick up on this fact by providing a link to search results that can be subscribed to. Bookmarking services can pick up the URLs directly from the address bar. Browsers that support OpenSearch will also show new functionality in the search bar. Additionally, repository tools that support the Open Archives Initiative Object Reuse and Exchange (OAI-ORE) can take advantage of the linked-data RDF/XML views. But most of all, members of the community who are able to parse the HTML code for the site will be able to provide new kinds of access back to the newspaper content available in Chronicling America.
The new website is built using the Python programming language, Django Web framework, RDFLib, Apache Solr search server, Apache Web server, and MySQL database engine.
See Chronicling America at http://chroniclingamerica.loc.gov .
Last Updated: 10/22/2012