Preserve OID mappings during import, and redirect failed OIDs?
Summary: This is a suggestion from Eric, to work around the problem of hardcoded OIDs both within the copied Space and elsewhere.
Original email from Eric:
Hardcoded OIDs, which don't work because every Thing has a new OID. I'm running into this a fair bit, since both of my Spaces use Computed Name, and when you use Computed Name the system kind of steers you into using OIDs to link to things unless you're savvy and careful enough to set up your Models and workflow to compensate for it. (URLs are OID-based, there's no Link Name shown in the page header, etc)
(This is especially thorny in situations where OIDs have been used in URLs outside the system - e.g., the Spirit Island FAQ, where there are hundreds if not thousands of OID-based hyperlinks to it elsewhere on the web. Even if there were a magic OID updater in the import process, it would only fix the OIDs inside of the Space itself.)
To avoid both the "pain" case of manually updating loads of in-Space OIDs and the "practically impossible" case of updating URLs elsewhere on the web, could Querki simply punt failed OID lookups to a lookup on a secondary OID-like property such as ImportedOID? (And ideally, set that property automatically during the Export-Import process, though worst-case I can set it manually prior to Export.) This has mild security implications - it could allow an attacker with read access to two Spaces to determine if Space A originated from Space B (by entering an OID known to exist in Space A and seeing if it results in seeing that Thing in Space B) - but even if that's troublesome, I feel like the benefits are well worth adding a single "Map old OIDs?" checkbox to the Import process, and the logic doesn't seem difficult. (Of course, I don't know Querki's internals, so I say that with a blithe obliviousness of possibly-critical details. But conceptually.)
