'tagme' and request tags

Alright. This is going to look like a royal pain in the derriere, but bare with me.

So far, we have these request tags:
translation request
source request
artist request
character request- only has one post, but could be useful if applied more often.

Now, looking at this, it's looking almost like our old nemesis tagme has been rendered obsolete. I have absolutely no problem with this. tagme is one of the most abused tags on danbooru, often misused as a catch-all for all the request tags and assraped by newbies who mistake it for a 'tag this for me' tag. I will even admit that as I newbie I thought this, because that seemed to be it's primary use.

Tagme has shrunk recently to only about 250 pages. I think this is a good opportunity to kill it, once and for all, and replace it with the following unambiguous system:

source request- Used somewhat like the current 'tagme'; for when an image is thought not to be original but the series isn't known.
artist request- When the artist is not known, or they are know but their name is not due to the language barrier.
character request- When the series(s) are known but the names of the characters are not.
name characters- Used just as it is now, as the tag for images with many characters the require tagging and notes.

The point would be to eliminate the ambiguous nature of tagme and get rid of the #1 excuse for lazy tagging. Of course, the key to this working would be cooperation between all of us. I'd suggest these steps:

-Remove tagme from all deleted images. There's no reason for that tag to be there if the image is for all intents and purposes gone.
-Look for images that have nothing but tagme or very few tags apart from it. Tag them and remove it, there's a good chance it was lazy tagging and nobody really wanted the source.
-Look for images that have their series, characters and artist all tagged and remove it from those.
-Look for images that have their series and characters tagged but not artist (and aren't tagged lazily) and switch it for artist request
-Look for images that appear original but have no artist tagged and switch it for artist request
-Look for images that have their series tagged but the characters not and swap it for character request.
-Once only source requests are left, mass-edit the rest of tagme images to source request and let it die when the server cleans out. Might do to make a public note somewhere that the tag is not longer in use.

Thoughts? Opinions? If I missed anything in my little plan please correct me, I've been awake for a little too long and my brain is a bit fuzzy.

Honestly "solving" the issue of tagme abuse will only occur by effectively banning the tag. That's not to say I'd support the idea. I see a huge flaw with your idea, which is that now instead of having one convenient tag to cover effectively the overall objective of those tags, you will have to search individually each tag.

Frankly it'd make more sense to have them all imply tagme, as that's the overall objective of those tags (they're just specialized versions of it). You're pretty much trying to remove the swimsuit tag by forcing everything tagged swimsuit into bikini, school_swimsuit, etc.

Also remember that tagme is also automatically applied to images uploaded with no tags. That behavior would need to be changed too.

The x_request tags are less intuitive than the tagme_character tagme_series (and what would be tagme_artist) system set up over on 3db, but I created the tagme_x system so maybe I'm not really impartial.

Just had a thought somewhat related to this discussion spurred by the "redundant tags" thread. What are people's thoughts on an image without a copy, artist, or char receiving the appropriate _request (or tagme_ if you roll that way) automatically? This assumes, of course, that there is much less original than non. Looking at the stats, this...appears to be the case (reduce 18924/345367).

jxh2154 said:
Also remember that tagme is also automatically applied to images uploaded with no tags. That behavior would need to be changed too.

Force users to include at least one tag that isn't tagme or *_request? Maybe include a notice about the *_request tags and how one can at least try to tag stuff like hair/eyes if nothing else...

at the very lesat the tagme tag should be removed once an image is properly (or at least somewhat) tagged.

This is a problem I recognize but I also don't really want to make any drastic changes like automatic tagging. In any case I've recorded most of the ideas presented here: http://trac.donmai.us/ticket/525

DschingisKhan said:
What are people's thoughts on an image without a copy, artist, or char receiving the appropriate _request (or tagme_ if you roll that way) automatically?

I can think of a couple corner cases here:

• black_rock_shooter is a copyright tag that doesn't have any related character tags. If the system added a character_request tag when no character tags were present then black_rock_shooter posts would erroneously be tagged with character_request.
• Some characters don't have copyrights (nevada-tan, longcat, barack_obama). Same problem as above, these shouldn't get automatically tagged with source_request because they don't have a source.

evazion said: I can think of a couple corner cases here:

Your examples are good ones, and they're why I'm adverse to making *_request tags automated.

DschingisKhan said:
This assumes, of course, that there is much less original than non. Looking at the stats, this...appears to be the case (reduce 18924/345367).

I suspect however that original is one of the most underused tags. Most users seem to just leave the character out rather than using original when they don't recognize a copyright.

sorry for the necrobump but i share the same sentiments of the OP. tagme had experienced an incredible surge last year thanks to the impressive tagging work of a single individual. it has grown too large (from something manageable) that i wonder if it's still useful for taggers now that we have every accurate and specific *_request tags that address any case imaginable.

populating tagme and some related tags (forum #82034) have become more as a nuisance/annoyance than being help (even if that's the intention). browsing the tag history has become a pain (almost impossible) looking for mistags/errors/vandals if one is greeted with not one, not two, not 10, but literally an ocean of tagmes and placeholders tags. it's even harder that this occurrence is more frequent (and more in magnitude) than jxh2154 himself when doing his regular admin run.

i strongly suspect that these tagmes and placeholders tags are inserted using very simple metatags which made them (already) obsolete a long time ago. if people still find relevance to the likes of tagme, i suggest monitoring/moderating them. or reserve tagme's only use as the default tag to save posts from obscurity. other than that nothing else.

why is there such an urgency to populate these tags just asking for help? isn't it more helpful to actually depopulate and replace them with more relevant tags? or should we all populate and retain them because the wikis said so?

I'll use this topic as we speak of requests. We have translation_check tag, but what about situations where one tag all characters present in uploaded art, but they're not sure of they did it correctly. I've tried to use character_check tag, but it seems that it's not used and after asking one janitor, I was told to share such doubts by posting them in a comment below. What do you think about it ?

Personally, I don't think a check_character tag is needed. If you're not sure about the characters, just use character_request without tagging them. If you tag them incorrectly it could end up confusing people.

On a related note, I also doubt the usefulness of check_translation. According to howto:translate:

Make sure you completely understand what they're saying. It's worse to have an image translated badly, than to not have it translated at all.

Unfortunately, I've seen a number of bad translations that just use the tag as an excuse.
Does anyone else think that use of check_translation should be discouraged, or does it actually help posts to be correctly translated?

check_translation usage should definitely be discouraged. If you're not confident in your translation you shouldn't be translating to begin with.

The times I've used check_translation mostly include when I'm reading a piece that comes across as nonsensical or the English side of the sentence structure is mangled to where I can't repair it without possibly wrecking the meaning. Unfortunately I think that might also mean that translators aren't too keen on checking the posts, as they may contain complex puns or things that make them difficult to translate in the first place. But, yes, I agree that the tag should never be used by a translator themselves.

This is a heads-up, but I've added a second paragraph to check_translation explaining that it's not a crutch for bad translators. I'm seeing the tag more and more frequently now, and it's clear it's almost exclusively used by people a month into Japanese 101. Train wrecks like post #1335028 are not acceptable.

Hillside_Moose said:
This is a heads-up, but I've added a second paragraph to check_translation explaining that it's not a crutch for bad translators. I'm seeing the tag more and more frequently now, and it's clear it's almost exclusively used by people a month into Japanese 101. Train wrecks like post #1335028 are not acceptable.

Good job, I hope this helps cut down on all the bad translations.

A related problem is people who don't even use the the check_translation tag most of the time, but just post a terrible translation and call it finished. The person who translated that post being one of the biggest offenders.

I think that people who consistently post bad translations need to be warned more often, in the form of mails or neutral records. So far I've never seen people getting records for poor translations, although I've seen a lot of poor translations themselves.