search

iftechfoundation / ifdb

The software behind the Interactive Fiction Database (IFDB)
Other
25 stars 18 forks source link

Categories #837

Open 6ru35l4y3r opened 3 years ago

6ru35l4y3r commented 3 years ago

This sort of covers iftechfoundation/ifdb#844, iftechfoundation/ifdb#843 and iftechfoundation/ifdb#841 in a broader range.

I want to suggest to replace the tag system by a category system as used in Wikipedia. Which is, in a nutshell, a tree diagram of tags, so a two-dimensional tag system, if so you want. There's unspecific higher categories and specific lower categories. Examples:

Advantages:

An example from Wikipedia: 1893 is in the category "2000s interactive fiction" which is in the category "Interactive fiction by decade" which is in the category "Interactive fiction". It is also in the category "Video games set in Chicago" which is in the category "Video games set in the United States by city" which is in the category "Video games set in the United States" etc etc etc - Very easy to navigate, and a tremendous step forward for a) the reader and b) analysis.

qdacsvx commented 3 years ago

A hierarchical category index could be valuable. Not everybody would find complex nested categories easy to navigate. Who would create the categories / index? How would you keep it up to date? If taggers don't like this complex proposal, who will add games to the index?

-1 People need shared public tags. Tags are simple. Everyone can tag. Tags are useful for informal categories and cross-cutting concerns. Tags are flexible and "speak the language of users".

It's not necessary to destroy tags to have an index. Cooperation is possible and beneficial. When a tag becomes popular, it could be added as a category in the index. That way the index will track user needs.

This seems like two suggestions.

  1. hierarchical category index
  2. remove public tags from IFDB
6ru35l4y3r commented 3 years ago

Who would create the categories / index? Formal categories such as authoring system, year, or genre should automatically be generated when creating a data record. Categories that demand knowledge of the game content would need to be created or suggested by the readers.

How would you keep it up to date? At this point, after tagging quite some games, I cannot think of a category that would become outdated. Same applies to tags.

If taggers don't like this complex proposal, who will add games to the index? The taggers. Those who like tagging will also like categorizing, It would appeal even more to their sense of order. ;-)

-1 People need shared public tags. Tags are simple. Everyone can tag. Tags are useful for informal categories and cross-cutting concerns. Tags are flexible and "speak the language of users". Letting everybody tag has led to distinctive tags such as "sci fi", "sci-fi", "scifi", "science fiction", "science-fiction" and "sciencefiction". I do not understand in how far tags "are useful for informal categories and cross-cutting concerns", or in how far they are flexible.

It's not necessary to destroy tags to have an index. Cooperation is possible and beneficial. When a tag becomes popular, it could be added as a category in the index. That way the index will track user needs. That's sounds good.

qdacsvx commented 3 years ago

Thank you for answering my impertinent questions. Now I understand your suggestion better.

Who would create the categories / index? Formal categories such as authoring system, year, or genre should automatically be generated when creating a data record. Categories that demand knowledge of the game content would need to be created or suggested by the readers.

I like the idea of auto data import. Currently bibliographic data e.g. blorb is embedded in game files. IFDB could have an importer so someone adding a game could simply select the game file for analysis. Or IFDB could recommend/link an external analysis tool. Currently all data is entered by hand. Except for bot additions.

How would you keep it up to date? At this point, after tagging quite some games, I cannot think of a category that would become outdated. Same applies to tags.

Tags can become out of date if nobody uses them any more. User tastes and needs change over time. Tags reflect that. Some old tags may be valuable to certain communities. But others are just zombies. Too many tags makes it harder for people to learn which tags are important/useful.

The difficulty with a directory is that it will be hard to keep adding new games to it without dumping them into 1-2 popular places resulting in too long lists of not very related games. This was the fate of most hierarchical web directories back in the day.

If taggers don't like this complex proposal, who will add games to the index? The taggers. Those who like tagging will also like categorizing, It would appeal even more to their sense of order. ;-)

It's easier to tag (1 choice) than to select a category (requires 3 choices for 3-deep hierarchy). Tags have no restrictions on civil content. But the value of a directory is the ability to make fine distinctions between close categories. But this value can only be sustained if great care is taken to ensure that only appropriate works are added to each leaf category. In other words, the directory owner will often say "no" to users who want to add their favourite game to their favourite category. This refusal is frustrating to users. But tags never say "no". I'm afraid users will convert to taggers to avoid the hassle.

-1 People need shared public tags. Tags are simple. Everyone can tag. Tags are useful for informal categories and cross-cutting concerns. Tags are flexible and "speak the language of users". Letting everybody tag has led to distinctive tags such as "sci fi", "sci-fi", "scifi", "science fiction", "science-fiction" and "sciencefiction". I do not understand in how far tags "are useful for informal categories and cross-cutting concerns", or in how far they are flexible.

Synonyms could be dealt with by moderation options to mark two tags as synonyms or to merge two similar tags iftechfoundation/ifdb#806 .

Informal categories include whatever is important to the writer. They don't have to get permission from a directory curator who often says "no, that's not a proper category!".

By cross-cutting concerns, I mean distinctive characteristics of a work e.g. a female protagonist, which are difficult to place in a hierarchy without e.g. including all tags as leafs at each last branch. But a directory shouldn't be large and random, like all tags, otherwise it is hard to find things!

I'm unwilling to support +1 the suggestion while it includes removal of tags. Tags can do things a good directory can't. And vice versa. You could propose that if the directory system becomes more popular than tags, tags could be retired. I would accept that.