Semiannual Report: May 2011
The Pleiades project has continued to build on the successes reported in our first semi-annual report (November 2010).
The following tasks were undertaken in accordance with the work plan during the period 1 November 2010 - 30 April 2011:
Rollout of Classical Atlas Project content: With the exception of roads, all Classical Atlas Project content has now been published in Pleiades. Most of this work was completed well ahead of schedule in November 2010. Approximately 200 places missed in the initial process (due to an idiosyncrasy in the formatting of the CAP data) were properly added in May 2011, shortly after the end of the current reporting period. The live, running total of ancient world resources (show on the home page of Pleiades), now stands at 31,953 places, 26,493 names and 32,729 locations.
As discussed in our previous semi-annual report, much of the technical infrastructure for roads and other linear features (like walls) is already in place. Our testing collaborators have been delayed in their project and so we, for the moment, wait to move on full import of the road features from the Classical Atlas Project. We would prefer to have a user-validated data model for these before fully populating. Two other external projects dealing with fortification walls are considering collaboration with Pleiades and they may provide concurrent testing opportunities in the coming year.
Collation and integration of content from the Digital Atlas of Roman and Medieval Civilization (DARMC): The DARMC team at Harvard succeeded in matching 11,655 of their data features with Pleiades places. These were published to Pleiades in March 2011, bringing the total number of precisely located places in Pleiades to 14,235. DARMC has also provided about 28,000 additional sets of coordinates not yet matched to Pleiades places. Some of these have no analogs in the CAP data, since they are later sites mapped in the Tabula Imperii Byzantini. These will therefore be entirely new to Pleiades. But we expect that a large number of coordinates for roughly located, existing Pleiades places will yet be obtained through another iteration of name matching, which we will conduct. Exploratory comparison of the data is underway to ascertain the range of techniques that will be necessary to identify doublets and match despite name variants and typographic errors. Exact timing of this work will be determined after community consultation.
Editorial oversight and content publication: Pleiades has published 925 place, name and location resources since our last semi-annual report. Editors have been active in checking and approving these changes, and in discussions with community members about adding additional time periods, language/script combinations and other features to Pleiades. We have also begun, where necessary, reaching out to external experts to help resolve matters that require local or specialist expertise.
Outreach and community building:
As we reported previously, the addition of blog-like news capabilities to the Pleiades site has made it easier to communicate frequently with our users and the general public.
Since its introduction in November, our “Frequently Asked Questions” section has grown significantly.
We published minutes from our first Pleiades “Community Time” report in December, 2010. Our next community time meeting will occur in June 2011.
An upgrade to Plone 3.3.5 provided much-needed fixes in our versioning system. Logged-in users can now review change histories for all content.
Sean Gillies represented Pleiades at the 2011 annual meeting of the American Association of Geographers, speaking in a special session on gazetteers. His AAG slides are available online.
We have recently added new site functionality to permit users and other authenticated visitors to comment on, and ask questions about, help items. We intend in coming months to extend this functionality to individual place, name and location resources in order to lower barriers to contribution of content.
Our proposal for a poster/demonstration session at the Digital Humanities 2011 conference has been accepted. Elliott and Gillies will be traveling to Palo Alto in June 2011 for this purpose.
User-driven upgrades and initiatives:
In January 2011, we rolled out a new way for users to get access to Pleiades content: nightly exports of data in tabular form. These data tables are in increasingly wide use by a number of collaborating projects in order to get up-to-date copies of all published content to support research tasks or cross-project linking (see further below). In April, a compressed KML export was added to the nightly export routine.
Over the course of the last 6 months, we have been refining a mechanism for scripting batch changes to arbitrary large collections of Pleiades content, taking data in similar tabular form as input. This mechanism was first used in January to add additional information from the Barrington Atlas Directories to over 20,000 Pleiades places. Subsequently, it has been used to add tags for “materials mined” to all the mine and quarry features recorded in Pleiades. A group of Pleiades users are currently working with Elliott to collate the contents of Current Archaeology in Turkey against Pleiades. The results of this work (primary reference citations but also many modern Turkish site names) will be added to Pleiades using this new method.
The identification of modern names for places is now supported in Pleiades with the same facilities for language/script identification, automatic transliteration and citation of reference resources as historical names. As of this date, the following language/script combinations are fully supported by Pleiades: Classical Greek (in both Greek and Latin script), Coptic, English, Latin (in both Latin and Greek script) and Turkish. Support for Arabic, Demotic, French, German and Italian languages is in preparation.
Layout, navigation and visualization have been improved through a series of incremental changes. These include the renaming of major browse paths, the reorganization of news and other portlets, the integration of a “recently modified resources” map on the homepage, and other user interface refinements all meant to speed access to content and provide better contextual awareness. Significant enhancements have also been made to the map interface in Pleiades (see further, below).
Pleiades describes more than 25,000 names and Plone (our application framework) provides little out of the box for navigating them. To try to improve the situation, we are experimenting with computing navigation facets, derived from Pleiades vocabulary terms. The new interface at http://pleiades.stoa.org/browse_names is linked to by several other pages on the site as well as by the central image in the scoreboard on the front page.
A number of separate projects that aim to reuse or link to Pleiades content have sprung up or advanced during the reporting period. We have been working closely with these projects to facilitate their work and refactor our data export mechanisms to support their needs. Key examples include:
- PELAGIOS: a multi-project consortium that is identifying best practices and normalizing approaches to using Pleiades identifiers in Linked Open Data applications. Two of the partners have recently reported substantive advances.
- Google Ancient Places (GAP): a Google-grant-funded effort to prototype the workflows needed to identify ancient places from unstructured texts (books) so that researchers can reference these ancient places in Linked Data applications.
- Regnum Francorum Online: via our collaboration with DARMC, RFO has been experimenting with visualizing Pleiades content (with links) in their map interface.
- CLAROS: The World of Art on the Semantic Web: In a first step, CLAROS has linked its place references to approximately 900 Pleiades place resources. This provides a mechanism, e.g., for determining what personal names in the Lexicon of Greek Personal Names are associated with a given ancient place.
Add Webcasts: We have begun working with interested members of our community to scope, script and produce new webcasts for user learning. We are currently discussing a step-by-step webcast designed for those interested in adding new and/or better coordinates to existing Pleiades places.
Archiving: As proposed, we have begun using the NYU Faculty Digital Archive to deposit static copies of Pleiades content with a third party for long-term archiving. Our first deposit was made in April, and consists of the full set of nightly tabular exports, with enhanced metadata. We will continue to make such deposits on a quarterly basis, and will include the full KML export as well. The inaugural deposit for 13 April 2011 may be found at http://hdl.handle.net/2451/29930.
The following tasks, originally planned for later in the period of performance, were undertaken early during the period 1 November 2010 - 30 April 2011:
Search improvements: Spatial criteria have been added to the advanced search and custom collections functions so that users can constrain discovery to particular geographic areas easily. We have also added the option of KML formatting for search results, enabling users to visualize their search results in Google Earth.
Map interface improvements. Over the course of the reporting period, Pleiades maps have undergone an iterative process of improvement. Maps are now significantly bigger and work for both published and draft content (they were previously available only for published content). Pop-ups in both the on-site maps at Pleiades and in the KML we produce for use in Google Earth have been retooled to better represent their origins in Pleiades and to make it easy for users to find their way back to the resources represented. The representation of roughly located places has been significantly refined. User frustration with our original bounding boxes has led us to an innovative graphical representation: “cloud” icons indicate the general proximity of groups of roughly located places. Clicking on these icons produces a nicely formatted list of links.
The following tasks, originally planned for the reporting period, were deferred to a later stage of the project:
Support for relationships: In order to accommodate needs expressed by the community (as outlined above), we have deferred further work on relationship modeling to summer 2011.
Zotero integration: We continue to defer intensive work on a bibliographic management tool in lieu of Zotero (see comments in previous semi-annual report). Significant work on bibliographic management is underway in the context of our partner project, “Integrating Digital Papyrology.” We will be taking stock of these developments and planning our approach during summer 2011.