Posted under Tags
BUR #59840 is pending approval.
mass update comic speech_bubble -> -speech_bubble
It's just padding.
Also this is a BUR mainly so people might pay attention to this topic. I am aware this is not a clean solution. A major problem is that people will continue tagging a generic-sounding tag such as speech bubble on comics "because it's right there", despite it being obviously padding.
Alixiron said in forum #440853:
It's extremely unintuitive and comics don't necessarily always have speech bubbles either.
We have silent comic for those cases.
I don't think it's padding because speech isn't always rendered within a bubble in comics. Right now, I can't even check how many posts that would entail because it times out. Additionally, there's plenty of stuff in comic speech_bubble that likely should not be tagged comic to begin with, but is because of the inverse problem, people seeing speech bubbles (and similar common comic features) and tagging it because they're tagging comic based on vibe and not what comics actually are (which gets even worse with the *koma tags, due to people assuming it's just "panels" and not a specific format, forum #363642, and even moreso with 1koma, topic #23117).
Additionally, preventing the use of speech bubble means speech bubble subtags would also be off-limits due to implying the tag (or otherwise needing to), like shared speech bubble, the various spoken_* tags, and even thought bubble.
Damian0358 said in forum #440856:
I don't think it's padding because speech isn't always rendered within a bubble in comics. Right now, I can't even check how many posts that would entail because it times out.
I am purposely ignoring the tehcnicalities of a BUR like this and the fact we can't stop people from tagging it with a name like this, but (and this is just a guess) I would assume speech bubble with borders is the "normal case". Normally we'd prefer to tag the exception and not the default case, so this would lead to something like borderless speech bubble.
ANON_TOKYO said in forum #440864:
Normally we'd prefer to tag the exception and not the default case, so this would lead to something like borderless speech bubble.
That's not a good name for that, and I don't mean in the "if it's borderless it can't be a bubble" sense, because you could just as well imagine speech where you just have the \/ bit next to the character's mouth (ex. post #9655851)... which I'm having difficulty finding because comic -speech_bubble has both completely unbordered text (post #7265276), untagged speech bubbles (post #5356871), partial speech bubbles (post #6067272), subtitles (post #7876291), speech bubbles without anything indicating character source (post #6474317), square speech bubbles (post #11406001), colored speech bubbles you wouldn't be able to tag because that tag implies speech bubble (post #8638966 is bordered, but you can imagine an unbordered example), etc...
There are 500k comic posts without the speech bubble tag and most of them should have it. Who is going to populate this? No one is, because it's a waste of time. This is what's frustrating about arguments like this: people say a tag is valuable and we should keep it, but no one is willing to populate it. It's saying "here, you waste hundreds of hours of your life on this, not me".
If we had any sense we'd tag the fact that speech or dialogue is present, not the bubble itself. It only takes a few seconds of thought to realize that speech can be in bubbles, boxes, or text outside of a bubble, and which is used isn't very meaningful. So even the argument that speech bubble can be used to find characters speaking doesn't work. It's another case of taggers focusing on the concrete object, because it's easy to tag, rather than on what the object represents.
evazion said in forum #440871:
If we had any sense we'd tag the fact that speech or dialogue is present, not the bubble itself. It only takes a few seconds of thought to realize that speech can be in bubbles, boxes, or text outside of a bubble, and which is used isn't very meaningful.
But then we run into the problem that speech is hard to identify when you don't know the language you're uploading, because how would you know that the text present is actually speech and not thoughts or narration? Thought bubbles aren't always used for thoughts, and narration isn't always in boxes either. Last time the idea of a speech tag had been discussed was back in 2017 with topic #13770, discussing talking, and people err'd on the idea.
I'll be honest: does anyone care if the image is just dialogue or just thoughts? In my opinion, the only significant difference is between "dialogue" in the broadest sense, text on objects like packaging held by characters, author's comments, and utility text (e.g., on gacha banners).
Dialogue: post #11409539, post #10713178 (narration), post #11407683
Object text: post #10707685, post #11411477
Author's coments: post #11405983
Utility text: post #11411582, post #11399988
てはいさ said in forum #440873:
I'll be honest: does anyone care if the image is just dialogue or just thoughts? In my opinion, the only significant difference is between "dialogue" in the broadest sense, text on objects like packaging held by characters, author's comments, and utility text (e.g., on gacha banners).
Our base text tags are used whenever a given language is represented with the exception of Japanese with its supermajority over the site. Translation request doesn't have anything narrowing down what specifically folks want translated, with that being effectively decided by the translator (meaning that, for instance, they can opt out of translating common sound effectseven if we have a List of SFX for translation reference, or unimportant text like clothes writing or billboards if they wanted to, because they assume the thing people care most about for translation is just the core text, whether it be dialogue, thoughts, narration, etc.).
Having something to distinguish the various kinds of text in an image allows people to not just search for specific kinds of comics, whether for reference or similar, but allows translators to pick and choose contexts which they want to translate. Until we do something the fact that we don't have narrower searches for comics specifically, topic #30094, reducing searching options isn't particularly helpful for anyone.
In that regard, evazion's argument for a speech tag has some weight. But the reason why I emphasize the point of 'not being able to recognize whether something is actually speech or not' is that this can eventually result in speech basically becoming a text umbrella tag, since one can assume that most text in posts will be dialogue, thoughts, narration, etc. and indirectly through that also be a Japanese text tag, since doing speech -*text would give you Japanese text.
