Fair enough, but there is a significant category of sites that you probably care about that do benefit from something like this: collaboration sites.
If you’re engaged in building something together over time (software, an organisation, a movement) then there is often a huge gap between when an idea is first floated and when it is actually completed.
Part of the point of using something like Discourse as a collaboration platform is actually to facilitate the continuity of discussion over long gaps, rather than constantly restarting discussions from scratch.
The nice thing about collaboration as a market sector is that it is naturally linked to paying customers ![]()
I’m puzzled you think it’s a big change. Look at the code of the top topics model it would seem to involve only touching the compute_top_score_for method. The line that is currently
if period == :all
// Get all the topics without filtering by topic.created
.....
// Build the query
DB.exec(sql, from: start_of(period))
would become something like
created_period = SiteSetting.top_topics_created_period
created_period = period if created_period == ''
if created_period == :all
// Get all the topics without filtering by topic.created
.....
// Build the query
DB.exec(sql, from: start_of(created_period))
I’m sure there would be a bit more too it than that, but it seemed like you thought the whole algorithm computation would need redoing.