...
| Criteria from The Document Repository Service | Status | Analysis of the Trust and Identity Document Repository Index | |
|---|---|---|---|
| 1 | The platform service must be designed to preserve the repository’s contents for as long as Internet2 has an interest in maintaining access to its documents. |  | (Internet2 has not formally specified a target for how long these documents should be preserved. Informally, though, it's been recognized that it should be possible to preserve these documents even if Internet2 ceases to exist.) Internet2's Confluence service is not a good choice for long-term preservation. While a geographically distant copy is maintained for disaster recovery, there is no provision for fixity to address "bit rot," nor are point-in-time backups maintained to recover from human or software error. |
| 2 | It must be possible to search documents based on their metadata. | | Confluence can perform full-text search, although there are not strong methods for limiting the scope of pages to be searched, so it is not possible to search only the repository. |
| 3 | It is desirable that it be possible to create document indexes based on metadata searches. (E.g., create an index containing all documents authored by the InCommon TAC, or all documents with a #SAML tag.) | | It is possible to create multiple indexes, but the filter criteria for each index cannot be aligned with specific metadata fields. |
| 4 | It is desirable that the platform support full-text search across all documents. | | The documents are stored as attachments, so they are not searchable. |
| 5 | When there are multiple versions of a document, the platform must provide clear indication of which version is current, and which have been deprecated. | | Metadata fields were established to address this. [Update 3/30/2018: The metadata fields do not provide clear indication, so this was change to a negative status from positive.] |
| 6 | The platform must support stable, unchanging URLs for documents, as well as a strategy for maintaining URL validity after a platform change. |  | A DOI prefix, 10.26869, was acquired for the repository, enabling the creation of unchanging URLs of the form http://doi.org/10.26869/TI.5.1. |
| 7 | It is desirable that documents’ URLs be readily mapped from the documents’ identifiers. | | The document's Repository ID (e.g., TI.5.1) is easily identifiable within its URL. |
| 8 | The platform must provide very high disaster recovery capabilities. | | Internet2 maintains a backup server geographically distant disaster recovery site. |
| 9 | The platform must support point in time backup and recovery, largely to recover from human or software errors. | | There are no provisions for point in time backup and recovery. |
| 10 | The platform must provide high availability and good responsiveness. | | Internet2's Confluence service provides high availability and good responsiveness. |
| 12 | The platform must support the following administrative functions: Upload documents. | | Confluence provides upload capability for attachments. The size is limited (tens of megabytes), but that has been sufficient for nearly all of our documents to date. |
| 13 | The platform must support the following administrative functions: Remove documents. | | Documents can be removed from the repository. It is not possible to recover them, however, if the removal was in error. |
| 14 | The platform must support the following administrative functions: Manage metadata for documents. | | Metadata is managed through normal wiki page editing. There is no automatic validation or enforcement of specific fields, so editors must be careful. |
| 15 | The platform must support the following administrative functions: Extend the types of metadata available for documents. | | This is possible but requires manual editing of all documents' metadata to ensure consistency across all documents. |
| 16 | The platform must support the following administrative functions: Manage multiple versions of a document with a stable URL for the current version. | -- | Multiple versions of documents are handled as multiple documents, each with its own metadata. There is no provision for a stable URL that maps to the current version as new documents become "current." In the "TI.<document>.<version> Repository ID, the one with the highest <version> is current. |
| 17 | It is desirable that the platform support automation of administrative workflows. | | Confluence is limited with respect to its automation capabilities. Support for WebDAV would improve the ability to automate tasks. |
| 18 | The platform must be accessible, compliant with WCAG and other applicable standards. | | Atlassian is committed to accessibility for Confluence, and provides a VPAT. Most applicable VPAT criteria, however, are marked "Supports with exceptions," and there are many unresolved JIRA issues linked as exceptions. |
| Criteria Discovered During Implementation | |||
19 | The platform must be capable of registering DOIs with CrossRef, based on metadata that has been entered in the repository. | | CrossRef provides a RESTful API for registering DOIs, but nothing exists for our Confluence repository to use it. |
...