There are two main reasons why I thought reviewing and marking solutions would be helpful: (a) it allows moderators to immediately identify unresolved topics on the Support category page, and (b) it allows users searching for existing answers to quickly identify topics with solutions.
However, I’ve identified two issues with marking topic solutions:
In some cases no single post is the full solution. Various people supply different pieces of the puzzle, which amounts to a solution. I pick the most complete answer as “the solution”. I worry this could be confusing to future readers, who see a partial solution marked as “the solution”. But there is no way to mark a full topic as “resolved” (that I know of).
There is no way to see who marked the solution: the original reporter or a moderator. There is a danger of a moderator marking a solution which they think is clearly correct, but might not actually work long-term for a user’s particular case. The user (or other users) might feel like the discussion has been “shut down” if a solution is marked, and then not contribute, which is counterproductive.
These aspects of marked solutions might be resolved by switching to a system in which only the original poster is allowed to mark a solution to their own question. In this case, perhaps moderators should switch to using internal flags to track unresolved topics in need of attention (i.e., instead of using the “solution” flag to mark resolved topics that don’t need attention).
This relies on the original poster to do some extra work. I’d expect that many will be willing to, but some will forget, and some answers will slip through the cracks.
First, I’ve enabled a feature that’s relatively new where you can filter the support categories (or any category where the Solutions plug in is enabled) by topics that are solved or not. At a minimum, this could help with some aspect of triage:
I agree it can be hard to piece together a solution from a conversation. In cases like this, could the CET staff write a succinct summary post containing the comprehensive solution (possibly with links to revised documentation as I think we should use the forum to inform documentation improvements), and then mark that post as the solution?
I still think that both moderators and the original poster should be able to mark solutions, as we currently do. I think its important for the observatory to reserve the right to curate the forum, and ensuring solutions are marked at the end of the day is part of that. I don’t think there will be an issue with shutting down follow-up if CET staff explicitly invite the original poster to ask more questions if the solution or summarized solution aren’t clear.
@ktl’s suggestion of a discourse flag is neat but I think it might be confusing to have two public ways of marking solutions. It sounds like it might be worth it to connect posts to internal ticketing software to help track support work and assign staff to topics, like https://zapier.com/apps/discourse/integrations/jira-service-desk
I agree that both moderators and the original topic creator should retain the ability to mark reply posts as “solutions”.
In order to avoid cases where a moderator marks a reply post as a solution because it seems to be one – but unbeknownst to them it didn’t actually work for the original topic creator – I’ve changed my approach to only mark a post as the solution when it’s very clear that it resolved the issue (meaning, the original poster confirmed it worked).
I also like @jsick’s idea of writing a small summary reply post in cases where the issue was clearly solved but the solution is distributed across multiple replies, and then marking that summary as the solution – I’m going to try that when I next find such a case.
For topics without very clear solutions, I’ll continue to post replies that solicit confirmation (or follow-up) from the original topic poster. In a few cases already this has lead to the original poster marking a solution.
Continued discussion and feedback on moderation techniques very welcome.
