So the summary algorithm assumes that staff members are better able to judge what is important on a topic?
For that to really work, the staff members better not be liking anything just because they like it, or because it made them laugh. Maybe discourse should warn staff members to only like post s which are most important to the topic. This could be done every few hundred likes, just to remind them that their likes have a heavier impact on the summary.</sarcasm>
On a more serious note: why was this weighting even put in place? It doesn’t really make sense to weight likes for just staff. Realistically, it would make more sense if the likes were weighted on a graduated scale. Something along the lines of:
- Level 0: weight factor 0.5
- Level 1: weight factor 0.75
- Level 2: weight factor 1
- Level 3: weight factor 1.5
- Level 4: weight factor 2
- Staff: weight factor 3
Of course, for that to really work, the applicable weight factor would have to be stored as part of the like. That way a user shifting between trust levels wouldn’t affect the summary of a topic.