What you said are possible, but manual tagging is still best. Consider mate in 4 where it would lead to smothered mate. It is possible that before smothered mate is finally delivered there are other motif that are involved such as pin, deflection, trapping the queen and others. Here manual tagging shines. I believe the priority is motif coverage rather than speed of tagging.
Join the team in,
lichess.org/team/puzzle-metadata-tagging-projectWe will discuss what is best on this project.
I don't agree that manual tagging shines here: it is slower and also easier to miss tags.
If you add tagging-functionality to the puzzle-generator code, there is no reason why it wouldn't come up with multiple tags in the example you provided. It's just a matter of properly defining the situations where you want a pin-, deflection- or trapping-tag. In the game analysis, starting from the puzzle position and for all correct solutions, it would then check for all intermediate positions and for all tags whether or not the tag-condition is fullfilled.
I have a better answer as to why this is not possible.
All puzzles have already been generated. They are not being generated any longer, and in order to do what you say would require regenerating all puzzles and starting over.
I disagree with the premise that on-site tagging would be bad.
Sure there are people that would miss-tag, but hopefully, the community as a whole, would not. For that there needs to be an option to un-tag. Basically, like a voting system but for tags. When the quotient of the voting reaches a certain number, the tag could be applied/un-applied.
There could also be a team of "moderators" who perhaps "block" the tags, in the sense that for a specific puzzle, tagging for a given motif is blocked.
For this to work, there would need to be a tactical motifs page, a page where the main motifs are explained, so users know what to tag and what not. This shouldn't be too hard to create.
Obviously, only pre-defined tags could be applied. I cannot go on a puzzle and tag with "idjsaoidas".
"the community as a whole, would not"
The part of the community who has to much time on their hands to bother modifying tags is an entirely different scope that is pretty far from your premise. Peer review is good, but cannot be achieved with the current proposed model, especially since they aren't getting anything out of it.
"There could also be a team of 'moderators'"
More workload, nope sorry. Generally a good model should leave the moderation work to the community within the submission process itself, or ensure no moderation is needed.
Thanks for replying.
I honestly think that the community would be interested in adding/removing tags from puzzles. And it's something that would not take much time, in my opinion. One would solve a puzzle, as one normally does, and after the puzzle ends, there would be the information about the tags. All one had to do is upvote or downvote given tags. Like in chesstempo.
By moderators I meant a different team of moderators, not the current one. Dunno, just an idea.
More ideas...
-The people allowed to tag should have x games under their belt.
-There should be y amount of tags before it becomes public.
-Include a flag button if something is mis-tagged.
-The tags should be predefined "offensive" "defensive" "avoid checkmate" "find checkmate" "positional advantage" "center control" etc...
-Users should be able to multi-select which tactics they want to train on. (Everyone who trains focuses on specific skills at a time).
-
