Keeping Data Alive Supporting reuse & repurposing of 3D data in the humanities


Evaluation: Research Outcomes for Data Publishing Workflows


This project successfully identified the basic steps needed for successful export of data and linked metadata from CityEngine for ingestion into a generic repository. The key steps in this process have been scripted in python, and these scripts are publicly available. Guidance for the creation of an appropriate repository is available through this report. A metadata schema appropriate to the procedural 3D models and widely in use in the heritage community has likewise been identified and applied.


While the key steps in the data publishing workflow have been established, important challenges remain. These relate to paradata schemas, flexibility in allowed file formats, and the lack of automated methods for tracking edits and participation in scenes. Additionally, while the creation of metadata as defined in this project is relatively straightforward, building on the EDM schema, the definition of a standard for paradata for procedural modelling and alignment with a standard remains a challenge. This is unsurprising, as paradata is likely to be more variable between projects and between researchers, and resists codification to a fixed standard. This project assessed the CIDOC CRM as a metadata standard, and while deemed overly complex for the metadata requirements for procedural modelling, may be a viable and appropriate solution for paradata, as its descriptive richness and event chaining will support more flexible descriptions of work processes. The use of the CIDOC CRM for paradata should be further explored in future work.

As discussed in detail above, this project’s system relies on the use of the current Collada (DAE) standard. Because the Collada standard is one of the more rapidly evolving standards, there is a risk that models made with future versions may not match the expected format of the system. Because the version of Collada exported from CityEngine is controlled by ESRI, and currently there is no selection or user-facing naming of the version, the project’s system might need to be updated in the future if CityEngine’s exported version of Collada changes.

The greatest challenge remaining is automation of the tracking of information about model modification and participation in scenes to facilitate ease of use of the system. In the current version, a user is responsible for manually entering this information as paradata. Manual entry of this information greatly increases the likelihood that rich paradata will not be captured. Automatic tracking of edits and parent-child relationship, and the export of this information into the json metadata and paradata sidecar file would be beneficial. In this case, manual paradata entry would be reserved for the explanation of why modelling choices were made.