Discussion sort of moved from Discord here: Discord
JAV have their own StashDB instance found at https://javstash.org already which is where most (quoting third-parties) big collectors contribute / update / etc. However, scenes are still updated on StashDB. StashDB does not allow scenes from OF and we have FansDB as an alternative.
The problem with this, is that it results in different metadata between the two, because there’s not really an agreed way of handling translations to English among other issues with metadata. p.s. I do not store JAV so if I’m incorrect in anything please say, this is from my limited understanding.
I propose that we do not allow JAV content on StashDB anymore, and redirect it to JAVStash.
However, this would result in all old scenes remaining; I would propose that over-time these scenes are removed, if there is a matching entry on JAVStash, but we keep the scene as-is and do not update the metadata for these scenes.
Here is my take on this - javstash primarily holds data in japanese. and although that is what the original source data gives, given an intl community, i am inclined towards storing jav data but english translation in stashdb. This english translation would be from r18.dev though since that is used most commonly and in several places. Maybe we can disallow Japanese data on stashdb for JAVs?
I wrote on this and it’s trash heap of consistency and was told I was clearly wrong and “the community” knows what is right without a single place it can be referenced.
I do not think it should be on StashDB since it’s creation and data source is directly counter to every other agreed upon standard.
I’m wondering how, as someone who doesn’t collect or I guess have interest in JAV, the existance of JAV entries (that you aren’t matching anyways) in the database is diminishing your user experience to the point where you feel that all the data that editors have spent years entering should just be deleted?
Anyways, JAVStash and Stashdb have different approaches to data. JAVStash scene entries are automated scrapes created without fingerprints. There are no tag aliases, either for english translations, identical meanings, or even spacing, making tags not really usable without heavy stash-side parenting, even if you can read them. And some JAV do have official (likely machine translated) english titles, which only have a place on Stashdb.
It’s not about my experience being shit, it’s about having a consistent set of data which doesn’t change depending on who submits said data.
If everyone who contributes to JAV consistently is using JAVStash with their own unwritten guidelines, and a few people who want English translations do their own thing, it causes a problem when attempting to find consistent data. Not saying this is happening, but it could.
It is also not about deleting anything, but about finding a better way or place to store and handle the data surrounding it.
I will say that I was surprised to be pulling Kanji aliases on tags into my stash, which seems silly. Perhaps preserving original titles and names for JAV porn and performers can make sense, but descriptive tags? Why not eight different Eastern European words for ‘busty’?
Surely you can describe the content of a scene you are uploading in the English language. If you’re linking to English tags you’ve already done so? I don’t get it.
And it sounds like a core objection to relying on JAVstash is that its data consists of inscrutable Japanese characters, but you then port the same Japanese into StashDB?
Okay, I see what you mean in the original post now. The misunderstanding is that JAVStash isn’t just StashDB for JAV in the same way that FansDB is essentially StashDB for amateur platforms. “Different metadata between the two” like tags, performer measurements, aliases, the language… are the point. They have different philosophies and there’s no reason to say that only one approach is valid.
It’s like if an AmateurStudioBox instance existed for people who didn’t believe in grouping amateur re-releases like StashDB does. You wouldn’t say that those scenes shouldn’t be allowed on StashDB due to inconsistency.
Popular studio tags are added to Stashdb so it’s more likely they’ll auto match on scene submission, you can’t guarantee that every editor is using the exact stashdb tag schema.
I don’t see the problem with having multiple overlapping projects with different rules.
Stash is not the only application that can use stashdb as a data source, xbvr can use stashdb for matching files as well as a data source for scenes.
One project that I have thought about but not gotten around to yet is to use the BOT queue on both instances to cross link the databases so add the javstash scene url to the stash scene and the other way around. A plugin could then look for scene urls and add that as a stashid to the scene like the plugin I have for adding performer stashids. Both databases tend to use r18.dev and the same studio codes so in most cases the databases should match.
