I hope discourse.com creates a preference to extract topic-chaff from the topic-wheat, leaving "my problem is," supporting data, and the solution, with nothing between or after. Even my "solution" has become unnecessary, as the original sketch now functions with a populated, but once-empty file. Forty posts to boil down to an "s" in the protocol isn't as bloated as 1.5k posts to "demystify" nasm but if bots can be permitted to bloat a topic with vague posts, why not use bots to trim topics to the necessary supporting information for a more rapid transfer of information.
If you see a tangential or off-topic discussion in a forum topic that is having a harmful effect, please flag the first post in that sub-discussion:
- Click the β icon ("privately flag this post for attention or send a personal message about it") at the bottom of the first post of the topic.
The "Thanks for keeping our community civil!" dialog will open. - Select the "Something Else" radio button in the dialog.
A text field will be added to the dialog. - Type a message in the text field suggesting the divergent posts be split to a dedicated topic.
- Click the "Message" button.
This will bring the situation to the attention of the moderation team, who are able to split topics as necessary.
There is some nuance to the decision regarding whether to split a topic:
We will be less inclined to conclude a split is necessary when the topic author has welcomed or initiated tangential discussion. Conversely, we will be more strict in the case where the topic author has expressed the desire to keep the topic tightly focused, or when it is clear that the topic author is especially prone to being harmed by off topic discussion (as is the case with a beginner already in a state of being overwhelmed by the technical complexity of their project).
We are less likely to make a split when the extent of the divergence is minor, such as a couple tangential posts that are unlikely to instigate further off topic discussion.
Yes, I try to use the flag... sparingly... really... and I saw no tangential posts...
But...
I dream of a "winnow bot" controlled by an on/off checkbox for "verbose" or "Problem and Solution." I will certainly read the chaff (for you-know-what) but also will like a tl;dr brief so I can evaluate the middle parts. Something for the Discourse bots to evaluate.
Discourse does provide a AI-powered topic summarization feature:
I generally try to be conservative regarding the addition of new optional "bells and whistles" to the forum's user interface.
I personally don't find AI summaries useful. My sentiment toward integration of AI features into Arduino Forum is that if the feature can be provided by an external service (e.g., ChatGPT et al.), then there is no point in cluttering up Arduino forum with an inferior (because we are not using frontier AI models) redundant capability.
I welcome the opinions of the community on the subject.
Here is what I got when I instructed Gemini 3.6 Flash to summarize The Arduino Networked Lamp :