I believe that tags should be used to classify, not to identify. This is one reason why tags would be removed from the system if they're not joining questions.
Things to consider about tagging are the usage patterns. I can see three usage patterns for tags:
- Looking at the tags on a question to help figure out the question (e.g. which civ game are they talking about?)
- Selecting questions based on tags (e.g. search results of a tag, or filtering out questions based on tags)
- Selecting appropriate tags for a question.
So, consider two different tagging schemes, for the proposed question:
A. 1 tag per game
[sid-meiers-civilization-2002],
[sid-meiers-civilization-2010]
B. No unique tag per game, but a unique collection (if known)
[civilization] [glenn-drover] [eagle-games] [2002],
[civilization] [kevin-wilson] [fantasy-slight] [2010]
How do these stack up on the three use cases?
1A. If I know the publication date of the game I'm familiar with, then I'll know that it matches or doesn't match the game I'm thinking of. Otherwise, this doesn't help me know.
1B. If I know any of the publisher, designer, date, etc. of the game, then I'll know if it matches or not the game I'm thinking of.
2A. I have only 1 tag to search or filter on, and I, again, must know the game based on date.
2B. I can search or filter on a wider range of games.
3A. I need to know the publication date of the game in order to tag the question. This may be difficult to obtain.
3B. To tag the question, I only give what I know about the game (designer, publisher, date, etc.) and it's also allowed to leave it ambiguous. In an ambiguous state it's still useful for searching and filtering and even disambiguation, even if not complete.