We have several tags that are hypernyms to other tags. Probably the most prominent are punctuation, tenses, prepositions and terminology all in our top 30 of tags. I want to question the current philosophy behind those tags and unify the approach.
First let's have a look at those four tags and the hyponym tags.
- punctuation
- tenses
- past-tense, present-tense
- All tenses
- terminology
- programming with synonym programming-terminology, science
- other sciences, and fields of terminology
- prepositions
We have several different approaches here.
1. Synonym approach (prepositions)
Community decision to create synonyms thereby merging everything into the hypernym and foregoing any sub-level distinctions. A good solution when the hyponym itself would be rather meaningless.
Pro: One-time solution: Create synonyms
Con: Loss of distinctions
2. Dual-tagging approach (punctuation, tenses?)
Questions are tagged with both the hypernym and the more specified hyponym, e.g. punctuation and comma or tenses and past-tense. There are two directions to this approach. If the hyponym is unknown or irrelevant the hypernym tag certainly suffices. The other way around the hypernym tag should always be added if it doesn't hit the five-tag limit.
Pro: Both fine distinction and grouping easily possible
Cons:
- 50:50 chance to search for correct tag if not consequently tagged.
- Consequent re-tagging necessary
3. Utter chaos (terminology, figures of speech, several others)
While terminology is technically a hypernym of a few dozen other tags there is no mention of it in any of the tag wikis, the tagging behavior is arbitrary and it's in its current shape not a helpful tag. Generally the terminology tag is all over the place. Another example is figures of speech, it's a hypernym of metaphors, simile, personification, hyperbole and synecdoche.
Possible Solution
Obviously the utter chaos is not really a viable approach. The all-synonymous approach means a loss of a lot of information. Currently we operate in the utter chaos with many tags, which makes searching and the general use of tags less viable. There are also a lot of questions affected which makes manual editing a total pain. Still the second approach should be our goal to make those tags more helpful.
Thus, I suggest the following steps:
- Community consensus on which tags to cluster according to approach two.
- Script-based addition of the hypernym tag to all hyponym-tagged questions
- Script-flag questions that would hit six tags.
- Regularly repeat from 2. (Preferably regular background job)
Essentially, I am suggesting a SE-technology backed way to create tag hypernyms - likely somewhere in the moderator area. If a regular system-backed solution is not possible more or less regular script based changes by moderators would be my second choice. Third and last choice would be completely community based tag additions.