Home Forums Plugins and API Tabbles API (V4): start here Reply To: Tabbles API (V4): start here

TarekJor
Member
Post count: 7

@Dan-Novak thanks for your comment

First

Tag-dependency
When entering a particular tag, a second tag is also added to the tagged resource.

this is related with nested/tags => tag.uri => example tag:clothing/shirt (/ is “level.separator”)

Aliases (aka tag.synonyms)
Tags that get replaced with other tags (tag point another tag, maybe master.tag and slave.tags)

Performance hit, on Danbooru(*1)

Maybe is “bad database design” of indexes and queries (Tabbles could use graph or NoSQL database instead of SQL, if this performance is an issue)

(*1) @Dan-Novak: Both aliases and implications in Danbooru are apparently a performance hit

About “controlled vocabulary”, great idea, I agree with you

About Hashing

That is one feature, but imagine a “cross platform” “cross reference” always mapping “same file” with same “tags” or maybe a “massive collaboration” based o hash.tags

The possibilities are infinite, and even mapping “urls” as hash not only files (this is other feature)

Tabbles database could handle “item” as

item.name (on filesystem for example) hash.MD5 or SHA-1(master.id) and then “tags.list” (any format)

tags.list => (tag,tag,tag).ordered.AZ tag could be simple “string” or “string/string” “string:string”

I don’t know “in deep” how Tabbles handle “SQL Database” but maybe (if lead developers, or Maurizio think it is useful), make tests for NoSQL (json or graph) databases (increase performance on relations.

This is not “hard” to implement, because you no are using “F#” and “C#” (functional programming, hybrid) to “it is more semantic-friendly” than “other approach”

BTW, Keep in touch @Dan-Novak, we could experiment with things 😉

Best Regards