search

MellonScholarlyCommunication / spec-eventlog

This document specifies the requirements for implementing and hosting a Artefact Lifecycle Event Log.
https://mellonscholarlycommunication.github.io/spec-eventlog/
0 stars 1 forks source link

Discovering Event Logs #3

Open Dexagod opened 3 years ago

Dexagod commented 3 years ago

We ran into a problem with the Event Log linking.

We can define links to an Event Log from any artefact resource stored on your or any other data pod, or from the researcher profile as an aggregate for the events of all their artefacts.

Starting from the profile, any event log defined on the profile can be found, and any event log defined on an artefact can be found by following the references to the artefacts defined in the type index, and then looking for event log links on these artefacts.

The reverse however is not necessarily true. On discovering an artefact, there is no straightforward linking to the researcher profile on which the event log can be discovered.

So my question is: is it required to link to the eventLog from any created artefact? Or do we want to include a fallback that you can look at the researcher profile to discover the event log with events for the artefact. And if the latter:

mielvds commented 3 years ago

Hi @Dexagod ,

So my question is: is it required to link to the eventLog from any created artefact?

I'd say no. It will be quite common that multiple actors contain information in their event logs about the same artefact. The artefact would potentionally need to include an uncountable number of links. Looking at this from the collector perspective, I think the most common routes are:

  1. starting from the author to the event log of the author and everything linked to the author; and
  2. starting from a Service Hub to the event log of the Service Hub and everything linked to the Service Hub; and
  3. starting from a repository of eventlogs, probably managed per domain (eg. computer science might have a repository of all notable service hubs)

Either way, I think it's safe to assume that a collector will always have a seed list of pods/service hubs to start with. In the end, the collector provide a subjective and incomplete view on the artefact lifecycle, which is at least better than what we have today.

Dexagod commented 3 years ago

Either way, I think it's safe to assume that a collector will always have a seed list of pods/service hubs to start with. In the end, the collector provide a subjective and incomplete view on the artefact lifecycle, which is at least better than what we have today.

This is indeed reasonable. Although I would add some fallbacks out of spec to follow publisher / author / creator links to look for local event information when coming across an artefact.