User discussion threads / feedback

[mdoering]

Develop a generic user discussion system that works similar to github issues.
Within a comment it should be possible to crosslink users (@mdoering), names (#Abies_alba) and other discussions (#63).

Threads should probably be scoped to a given dataset in the Clearinghouse and default to the CoL and its extended version.

Develop name pages that exist for every name (in CoL?), tracking all data changes and any user comment from issue threads that mention a given name.

Provide a user dashboard that shows the latest notifications and a system wide summary for the latest activities.

[mdoering]

Hey @mdoering what do you think about notifying users like that?

[mjy]

Sounds like a kinda pointless idea at first glance. Why rebuild this? There is a meme that says all software evolves to adding an email client. Seems like Wiki's + GitHub/Lab integration + Discourse or existing forms in some combination could be the solution.

[mdoering]

Matt, I have thought about that a lot. Generally you are very right, this is basically a generic thing one should be outsourcing and reuse whats there. But it's the ease of use and integration that gets my attention. Why did github not use Discourse? Why does iNaturalist close all (Discourse) forum discussions about specific taxa and ask users to use their custom flag system instead? The added value is the tight integration and intelligent markup that allows to refer to your users, names and other things of interest. That provide notifications. Having a chronological thread about a name that not only includes comments from forum users. We certainly do not want taxonomic users to require to have a github account. Does TaxonWorks use Discourse? GBIF uses github issues for catching feedback. This works for technical users and sinply capturing a feedback. But it rarely ever leads to discussions.

I honestly dont feel it is such a big thing to develop, but sth that if done right gives a lot of advantage and helps building and binding the community.

[mjy]

Hi Markus. I generally concur, like I said, first glance. Part of my negative reaction is that this request comes up from many users, without their thinking about how it would actually help in a practical manner, i.e. how would things improve the data. I think this happens because it's easy to imagine. Indeed- TaxonWorks technically has most (all?) of this linked in this case to specific data, minus native cross links (we really need a generally agreed on Markdown extension for that stuff). What we find is most useful is the categorized annotations, i.e. those that mean something specific (Confidence/Data quality), rather than the general notes- which are nice for historical review, but poor for actionable decisions that improve the data.

You mentioned Github issues specifically in the initial post, which made me think it was for similar tasks. Indeed the iNaturalist post is pretty much a Github issue: "Closing to focus discussion about specific iNat taxa on the flags, not the forum.".

OpenTree does something nice where they embed Github issues in their app, so everythign is tracked in a git repo, but appears to users they are working natively. Best of both worlds.

[dhobern]

I think the obvious potential benefit that can/should come from this is the way that it COULD enable taxonomic researchers to leave information anchored to scientific names, species concepts, specimens, etc. for subsequent handling by whoever revises these taxa. At present, if I see a possible issue with the interpretation of taxonomic literature or in deriving species concepts and if I have no time, opportunity or incentive to do all the research and publish the findings, there is nowhere for me safely to leave these notes. A system anchored in CoL could serve this purpose.

[yroskov]

Please remember that CoL does not build and curate neither taxonomic concepts nor names. CoL only publish a view which originally built and curated in various database systems around the world. Any kind of dashboard on data aggregator side very soon ends with very messy information (as an example, have a look on recent comments in GBIF backbone github: https://github.com/gbif/portal-feedback/issues, and particularly, gbif/portal-feedback#1677). Dashboard might be useful on data provider side (at ITIS, WoRMS, TaxonWorks, Scarabs, World Plants, etc.), or on aggregator side, when all databases will be built and curated in a single IT environment.

[dhobern]

Thanks, Yuri - this all certainly depends on CoL proceeding further in work with the Commissions and with taking responsibility for handling names/concepts as a properly interconnected and stable infrastructure. The issues you identify with the GBIF backbone are primarily a result of our not having a stable version we are all improving together. Such improvements do not necessarily require all databases to be centralised in one system, provided that appropriate best practices are supported. Annotations on every individual database, with separate logins for every one, has (not surprisingly) proven to be a deeply flawed model. This is one role that a CoL can deliver as a consolidated brokering service.