The website has four distinct technical aspects: design and implementation of a database-backed website, conversion of the descriptive catalogue and bibliography, automatic processing of images, and overall design, styling and presentation of the website as a whole.
The project had previously created a Microsoft Access database which the InfoDev team was able to convert, improve, and import into an online MySQL database. Using the CakePHP framework meant that certain aspects, such as the creation of the administration section and the access control lists for who is able to edit the underlying database records, were straightforward to implement. The website uses an MVC (model, view, controller) paradigm to organise the application into input, processing and output sections as this facilitates reuse of programming code and faster application deployment times by handling tasks discretely. This paradigm enables easier production of different output scenarios because the interface is separated from the logic and data access and more straightforward migration between different data storage media. However, this approach does require careful initial consideration as to the separation of the application elements, but then facilitates agile changes when requirements shift during the course of the project.
Catalogue and Bibliography
The English translation of Corsi's descriptive catalogue was edited as a Microsoft Word document using a set of previously agreed and carefully applied styles to mark up various categories of information that the catalogue contains. By using an experimental documentation transformation RESTful web service (OxGarage, created by InfoDev;), the catalogue and its stylistic markup were able to be converted to more semantically descriptive TEI P5 XML (an international standard for textual representation;). Further bespoke up-conversion (using XSLT) introduced additional markup to allow easy interaction with the other aspects of the website. Because of this from any entry in the translated catalogue, users can click through to the database record for that stone, a large image of it, or an image of the original Italian entry. Editorial commentary provided by the project can be hidden by users should they desire. The project's bibliography and glossary also underwent a similar transformation enabling the items it contains to be linked to from the rest of the website. The InfoDev team provided an afternoon’s training in TEI P5 XML and the oXygen XML Editor to enable them to make ongoing corrections and updates.
The images of Corsi’s 1,000 polished stone slabs are one of the most important aspects of the site as it allows users to see the stones and potentially identify their own samples by comparison. Users can see thumbnail, image gallery, or individual stone views. The images were captured at 600dpi for archival purposes on a Canon 8400F flatbed scanner apart from a few translucent stones that required conventional photography. The InfoDev team undertook an automatic cropping of these images (based on an average stone size) using ImageMagick, as well as scripting conversions of the result to different sizes and cropping details for use on the website. The stones sometimes have multiple images (for back or side views where these are considered diagnostically important) which can be viewed as part of the database record for that stone. The pages of Corsi’s original catalogue were scanned, and the project stitched together sections from multiple pages where this covered a single entry. These were also processed as above to make them available in a number of sizes.