Any plans to host this on perio.do?

[periodo]

Any plans to host this on perio.do?

[rybesh]

In the short-term, it is not so expensive, but making a long-term commitment to keeping a public reconciliation endpoint active can be an expensive proposition.

Another reason we prefer to take this approach, though, is that we have some users who want different approaches to how reconciliation works (for example some need high precision while others want high recall), and this way they can run their local reconciliation server with different parameters to configure that (rather than us having to host multiple differently-configured servers).

[rybesh]

Thanks!

This reconciler is intended to be run locally (e.g. on localhost). We don't host it publicly because it is somewhat resource-intensive, and we are a small academic project with no current funding. Running the reconciler locally also allows users to reconcile against non-public datasets (e.g. ones that are still under development). This is in keeping with the general approach we've taken with this project (see our paper A sharing-oriented design strategy for networked knowledge organization systems).

[wetneb]

Ok, it's a shame! We (the OpenRefine project) should really consider running a server to host reconciliation services like yours (I cannot believe it can be that expensive for such services).