Small-ish issue with digital object identifiers

I’m no expert in this space, but I came across an issue with digital object identifiers because of my annoyingly persistent use (overuse? hahaha) of figshare. What happens if the archive tool you use for your data switches from one permanent link system to another?

Back in the early days of figshare, they used the handle system to provide a permanent link for data stored in their system. At some point they switched to using the DOI system. I have no idea when it happened and I don’t even think I noticed the change. The only thing I know now is that my older figshare datasets are full of dead links.

The point of using a permanent link, ie a handle or a DOI, is to maintain a connection to the source if the URL or data at that source changes. Any changes will result in a change to the metadata which will allow the permanent link to point to the correct location. This allows you to change the URL for a dataset on figshare, for instance, and the DOI link will point you to the updated location.

In my case old projects that were linked via the handle system are all updated with DOIs. Since the two systems are different, I have the unique situation of having broken permanent links! Obviously, this defeats the purpose of a permanent link. So it seems I have some work to do to find all the outdated figshare sets and update them, which presents a very tedious set of challenges.

Has anyone ever experienced anything like this? I’m not familiar with the internal workings of permanent link systems, but is there a way to easily move from one system to another? Does this present an issue for the future of web science where DOIs or handles are obsolete? I imagine in that world there would need to be a system wide effort to ensure everything is upgraded properly (like switching from paper to electronic records).