Fibery - Discourse Integration quick preview

Discourse Integration is here. Finally, we can accumulate feedback from all sources into Fibery. Here is the quick preview:


It’s gorgeous :fire: And I see my name! I’m famous :smirk:

1 Like

So good to see this!

Is there any way to actually pull additional data like the number of “hearts”? Or is there some other non-manual way you are calculating the “support” or desire for a mentioned feature from Discourse?

Hopefully it is not just viewed as a single “vote” per-thread. I imagine it could be complicated to reasonably calculate a “feature desirability” other than just by referencing the “hearts” on the first post that requests it. But avoiding manual tallying of desirability metric from forum sources would really be ideal…

Maybe in the distant future you plug-in to a machine learning API and can do sentiment analysis. :wink:

1 Like

We do have Like Count here, but not sure how it use it properly. Maybe we will just add references + likes and it will do the trick.

1 Like

Ah, good. Are the likes total/cumulative for the entire thread? I am glad you are thinking about all this stuff.

Looks like still some stuff to add, sub-categories for example (they seem to show as no Category for now), and tag import too would be great.

Also I’m wondering how the content part of it works. I have setup a test in one of my Fibery instances and it imported all content-to-date. But what happens if I e.g. connect some content to Entities or whatever, and then re-sync and a thread is updated? Does its contents just get appended to the end of the entity for that synced topic? Or perhaps the more challenging case, I can apparently edit the text in the rich text import, so do those edits persist after syncs? Or what if someone edits their post in Discourse? This seems a bit challenging to handle, so I understand if it’s not “perfect”. :smiley:

Also a small thing, but it would be nice if you could convert the @mentions as well as topic reply name instances into live links to Discourse Users imported into Fibery (if that option is chosen during integration setup). A minor convenience because of course there is a list of users at the bottom, but IMO would be a time saver…

Anyway, already this is pretty great, thanks!

I don’t plan on using this integration, but you raise a good point as I’m mainly hopefully that since we aren’t going to get any formal voting on features, that the “hearts” will be counted by the actual Fibery team to move along feature development. I don’t mean to beat a dead horse as unfortunately the external voting board was rejected, but I think you are getting at a serious flaw in Discourse: How can you be sure all “hearts” in a thread are for the feature actually requested, or rather for a particular comment that users liked within the thread. This could lead to inaccurate counts and analysis of what users are actually trying to upvote.


1 Like

Well, it’s not necessarily a flaw in Discourse fundamentally as it is a deficiency for this particular purpose. Which is why plugins exist for Discourse to handle the voting more formally.

That said I would tend to view Hearts on the 1st post most clearly as being a vote for the feature. If I were wanting to come up with a way to try to semi-accurately represent feature importance to the community with the built-in Discourse functions, I would probably start with first-post-hearts, and then somehow factor-in the amount of discussion generated as a multiplier or something. Of course that is not totally accurate either as the discussion may ultimately end in a “feature not needed” conclusion, so you’d still need to review and indicate some status, like “closed/resolved” for example on discussions that ended without a valid feature request. Still, for those discussions you left or marked as “under consideration” or whatever, the first-post-hearts x amount-of-replies metric would probably be more useful than just hearts alone.

1 Like

Yes you are 100% correct, that is how I approach Discourse as well - the initial request’s hearts are what I’d treat as actual votes for the feature. You bring up I have to say a bunch of arguments why the voting is a much superior solution to trying to decipher real intent of users within the Discourse threads. For example, if Michael jumps into a thread, he tends to generate a bunch of hearts, often more than the initial post that you and I are agreeing should be where you’d expect to find the actual upvotes. And Discourse is set up - I agree - as most apps in this day in age to allow actual posts to get their own “heart” progress. So it really doesn’t lend itself without the voting to measuring feature upvotes! Unless you are really reading closely and analyzing internally, which the Fibery team may be doing…but I’m almost certain the reason Asana, Airtable, who all have huge discourse forums, added the voting, because trying to make sense of the real intent of users with the native Discourse hearts is not very scientific.

So that makes me wonder what use you can get out of this integration if it tallies hearts on a Thread-level as you are discussing above? Unless you want to put in the manual work to analyze and decipher real intent, subjectively, it won’t be accurate.

Anyway, I think this is just another illustration of how useful the voting board could be, and how fringe this integration is for real feature prioritization. Don’t mean to sound harsh, but I think there is a lot of merit to what I’m saying if you look at the facts. Plus, the Fibery team encourages users to speak freely with feedback, so this is some of that :slight_smile:

Thanks as always for chiming in Oshyan!

1 Like

Yes, those are all fair points. I agree that having the actual feature vote plugin implemented would be superior. But I do think that a first-post-hearts count is useful, if not ideal, and so that’s what I’m sort of advocating for here in my above responses given that the vote plugin is not an option at this point.

1 Like

Agree with you again! If you had to figure out a way to “harvest” the hearts, it should be the ones on the first post (sorry to Michael but his highly-hearted follow-ups shouldn’t count :slight_smile:!

1 Like