We're considering both, but mainly listening to the "community voice" right now :)
Linking to data on other sites (including archive.org) seems like a better way to go, than pulling everything in (as you mentioned, storage reqs become a challenge fairly quickly)
We really had to step on the gas with this one. There has been just so much reverberations in the photo community that it became clear that it needs to be done (and quick).
This is a great effort! But yes, manual adding is slow, so the way is likely to crowdsource/automate it.
