🕸📝Fergus Duniho wrote on Thu, Jul 5, 2018 01:26 PM UTC:
Bringing together the two things I was getting at, categories should be exhaustive and mutually excusive, but tags don't have to be. A book I'm reading on SEO points out that categories should be used for important parts of site structure. When this site began, it was all HTML pages, and each page had to go into a particular directory. Given that we had to place each page somewhere, categorization made sense. But now that we have lots of content in the database instead of in HTML files, finding a particular place for each page is no longer as much of an issue. So, instead of organizing the site around the site structure, we have the opportunity to organize it around the nature of Chess variants. It's hard to come up with a strict categorization of Chess variants unless we make it about just one feature. For example, Chess variants could be categorized by their number of dimensions. If we wanted to categorize by cell shape, as well, we may start to have problems. A 3D game could use hexagonal cells, for example. If we use categories, what we find is that various Chess variants will fall into a variety of different categories, and some that fall into some of the same categories will also fall into different categories. There seems to be no easy way to sort or arrange Chess variants by categories. Instead of falling into strict categories, Chess variants differ from Chess to varying degrees and in different ways. So, it seems more natural to organize Chess variants around the ways they differ from Chess and around some other salient features than it does to try to strictly categorize them all. A tagging system works better for this than a categorization system. What I propose, then, is to use the existing categories for automatically generating some tags, then use tags rather than categories.
The main drawback I see to this is that categories are stored in the database much more efficiently than tags are. To find out the categories for a page, we just have to check one column in one row, but to find out the tags for a page, we have to find every row for that game in the Tags table. Additionally, if we allow public tagging, this increases the number of rows for each page. However, it isn't necessary to search the whole table, since the use of indexes narrows the search to just the rows we need to find.
One option is to repurpose some of the categories and add new ones, using them to note various ways in which games differ from Chess. This would make categories the main way to organize the variants, though they wouldn't work like strict categories. This would be more efficient and uniform than using tags, but the use of tags could give people more freedom to organize variants in ways that make sense to them. For example, people might want to use tags for family relations between games, or inventors might want to organize their games around different categories than the ones we've chosen to use. So, while tagging is less efficient and public tagging will lack uniformity, it does offer some advantages.
I'm not sure how we could use both categories and tags to organize the site. It seems like there would be a lot of overlap between how each got used, and there doesn't seem to be need to have both. Some of the advantages of tagging could be had by reworking the categories, but switching to a tagging system could offer other advantages besides these, though there would also be a cost in performance.
Bringing together the two things I was getting at, categories should be exhaustive and mutually excusive, but tags don't have to be. A book I'm reading on SEO points out that categories should be used for important parts of site structure. When this site began, it was all HTML pages, and each page had to go into a particular directory. Given that we had to place each page somewhere, categorization made sense. But now that we have lots of content in the database instead of in HTML files, finding a particular place for each page is no longer as much of an issue. So, instead of organizing the site around the site structure, we have the opportunity to organize it around the nature of Chess variants. It's hard to come up with a strict categorization of Chess variants unless we make it about just one feature. For example, Chess variants could be categorized by their number of dimensions. If we wanted to categorize by cell shape, as well, we may start to have problems. A 3D game could use hexagonal cells, for example. If we use categories, what we find is that various Chess variants will fall into a variety of different categories, and some that fall into some of the same categories will also fall into different categories. There seems to be no easy way to sort or arrange Chess variants by categories. Instead of falling into strict categories, Chess variants differ from Chess to varying degrees and in different ways. So, it seems more natural to organize Chess variants around the ways they differ from Chess and around some other salient features than it does to try to strictly categorize them all. A tagging system works better for this than a categorization system. What I propose, then, is to use the existing categories for automatically generating some tags, then use tags rather than categories.
The main drawback I see to this is that categories are stored in the database much more efficiently than tags are. To find out the categories for a page, we just have to check one column in one row, but to find out the tags for a page, we have to find every row for that game in the Tags table. Additionally, if we allow public tagging, this increases the number of rows for each page. However, it isn't necessary to search the whole table, since the use of indexes narrows the search to just the rows we need to find.
One option is to repurpose some of the categories and add new ones, using them to note various ways in which games differ from Chess. This would make categories the main way to organize the variants, though they wouldn't work like strict categories. This would be more efficient and uniform than using tags, but the use of tags could give people more freedom to organize variants in ways that make sense to them. For example, people might want to use tags for family relations between games, or inventors might want to organize their games around different categories than the ones we've chosen to use. So, while tagging is less efficient and public tagging will lack uniformity, it does offer some advantages.
I'm not sure how we could use both categories and tags to organize the site. It seems like there would be a lot of overlap between how each got used, and there doesn't seem to be need to have both. Some of the advantages of tagging could be had by reworking the categories, but switching to a tagging system could offer other advantages besides these, though there would also be a cost in performance.