Peekaboo said:
All makt åt TWYS, vår befriare![/i]Hooray for old, obscure non-English references.
Talar du Svensk, eller är du bara citera något slumpmässigt?
(Förlåt, jag fortfarande att lära sig språket)
Updated by anonymous
Posted under Off Topic
Peekaboo said:
All makt åt TWYS, vår befriare![/i]Hooray for old, obscure non-English references.
Talar du Svensk, eller är du bara citera något slumpmässigt?
(Förlåt, jag fortfarande att lära sig språket)
Updated by anonymous
Aurali said:
My dumb tail actually thought this would actual garner more opinions other then "TWYS is flawless", well, because it's not.
Ha ha ha... not.
I think that TWYS rule is deeply flawed, and should be changed as soon as something better would be found. The problem is that nothing better have been proposed yet, and especially your proposal with cumulating all possible tags failed to convince me. It would ruin almost all searches that are possible right now, and I don't see any pros that could compensate that.
So good luck in fighting with straw man. I'm out of here.
Updated by anonymous
Aurali said:
and no term for dick, no boobs? hmm.. using the examples above, I'm sure that dickboy is the only real term for this strange unheard of phenomenon. :p and what's a female? that term doesn't really fit now that we have the other examples. I propose cuntgirl, and for herm.. dickcuntgirl.
Aurali, this isn't funny anymore.
Updated by anonymous
Gilda_The_Gryphon said:
Ha ha ha... not.
I think that TWYS rule is deeply flawed, and should be changed as soon as something better would be found. The problem is that nothing better have been proposed yet, and especially your proposal with cumulating all possible tags failed to convince me. It would ruin almost all searches that are possible right now, and I don't see any pros that could compensate that.So good luck in fighting with straw man. I'm out of here.
What don't you understand, how would it ruin search?
Updated by anonymous
Aurali said:
and no term for dick, no boobs? hmm.. using the examples above, I'm sure that dickboy is the only real term for this strange unheard of phenomenon. :p and what's a female? that term doesn't really fit now that we have the other examples. I propose cuntgirl, and for herm.. dickcuntgirl. and I was talking about more then gender hun. anything that is defined in what other tags there are seems.. pointless
*facepaws* I think you out of all people would know what a 'male' is.
Updated by anonymous
Aurali said:
What don't you understand, how would it ruin search?
I think the most pressing question is, how would you implement this idea?
What are the things that stay the same, what things change, what things fall away?
What do the gain at what cost?
Like I already wrote you, a new/fresh booru with your proposed system for us to playtest would probably convince us more and easier than any words you or someone else could ever write.
abadbird said:
-snip-
May I rent you to check logic in various texts?
Updated by anonymous
NotMeNotYou said:
I think the most pressing question is, how would you implement this idea?
What are the things that stay the same, what things change, what things fall away?
What do the gain at what cost?Like I already wrote you, a new/fresh booru with your proposed system for us to playtest would probably convince us more and easier than any words you or someone else could ever write.
I will once We get down the line to the point of adding groups of tags to searches. If I still think it's a good idea anyway. .
Conker said:
*facepaws* I think you out of all people would know what a 'male' is.
because I'm a lesbian? .-. how does that make sense.
Updated by anonymous
Aurali said:
because I'm a lesbian? .-. how does that make sense.
Well what did you think the strapon was modeled after? A toaster oven? :V
Updated by anonymous
Conker said:
Well what did you think the strapon was modeled after? A toaster oven? :V
well.. it would give a girl a penis.. so a dickgirl .-.
Updated by anonymous
First of all, this proposal needs clearer and likely more thorough explanation before it gains productive traction. I believe the proposal does have merit and engages a legitimate issue, but the original post is somewhat sporadic. I present my analysis below (because I felt like it).
Fuzzy should imply that an attribute (read: tag) is open to interpretation whereas a static attribute adheres to a universally accepted definition. Perhaps a static tag can be thought of as a component whereas a fuzzy tag represents a conceptual arrangement of those components. What components should return a human, scalie, or herm in a search?
Anyways, I take away the following from the original post:
1. Fuzzy tags are currently treated the same as static tags. This causes fundamental disagreements over the qualifying parameters of fuzzy tags.
2. e621's search logic likewise does not distinguish between fuzzy and static search terms. Assuming all posts are tagged consistently, this requires users of the search feature to conform to e621's methodologies and rationales for handling fuzzy tags. History has shown that many guests are unable or unwilling to accept these rationales. This calls into question the intuitiveness of the current system.
3. Fuzzy tags cannot be explicitly defined. If they could, then they would be static. A fundamental incompatibility.
4. The two dimensional nature of still images generally results in incomplete information, unless a single picture contains multiple angles and perspectives of a scene. Certain artistic representations will lack key information present in other renditions of the same subject. Thus, the TWYS rule ignores or conflicts with attributes that are strongly implied but not explicitly illustrated. (This point was, I believe, intentionally redundant/circular to illustrate the flaw of TWYS.)
5. The proposal: Fuzzy tags/search results should be cumulative rather than subtractive. Currently, a user might search for one fuzzy tag and additional static tags to filter search results. I believe this is subtractive logic. In contrast, cumulative logic would check an image for the presence of certain static tags to determine the probability that the image contains a fuzzy tag. Now, I initially interpreted that this proposal was intended for programming behavior, but perhaps it was intended for user-defined tags.
Aurali also attempted to draw attention toward another flaw in TWYS's current implementation. I perceive the flaw in abstract terms... How do people handle complex issues? They break them down into smaller parts. Complex pictures and tagging schemes are often subdivided into smaller, specific tags. Likewise, controversial tags are often discussed from multiple constrained perspectives to attempt to identify and reconcile their controversies.
This strategy is scientifically proven effective when people work in groups (teams, really), but when people are left to their own devices? An individual's quality is often inferior to the group's effort. Users at all privileges struggle to interpret the recorded results of previous tagging discussions, invariably making subjective errors. e621's handling of fuzzy tags, specifically, is too bulky and nuanced to avoid these errors. Despite best efforts and intentions, strict adherence to TWYS rules is, in effect, unmanageable because even conscientious individuals fall victim to their own fallible subjectivity.
Thank you for this thorough analysis of what is being discussed. I find it particularly troubling that so few are willing to take a step back and consider "is this the best implementation of our goal? Are there flaws in it? If so, what are they? What can we do to improve this?" which are some of the fundamentals of scientific and critical thinking.
TWYS answers: What is physically present in the image? (penis, breasts, purple shirts, cybernetic limbs, cars, etc)
What's left unanswered is what subjects are being searched for? (hermaphrodites, lesbians, mind control, etc)
Searching for penises is easy. We can tell right away what the user wants to see. They want a rain of dicks to be showered upon their monitor. This involves purely physical properties.
Searching for lesbians is moderately difficult. Is two girls holding hands what they want to see? Do they want to see kissing? Do they want to see a longing look over a table? Do they want to see pussies grinding? We have a lot of physical properties here (especially here on e621 where there is so much sexual content as opposed to a more general search engine) but there are still these unseen properties. Emotional connection, relationships between characters, implied physical interaction, and other edge cases. How do we optimize this search and actually deliver to the user the content that they want?
Searching for incest is hard. Is presenting the user with pictures of characters saying "Oh sister~" what they want to see? Do they also want to be shown related characters with no indication in the picture that they are related but actually are? This would appeal to some users but not others because this is an entirely psychological concept and so the appeal of it depends upon knowledge of non-physical factors.
Another way to phrase the question that this debate revolves around would be: What do we want to be searchable?
These problems -can- be solved, it's merely a question of effort. For such things as post #326986 we have physical factors and implied factors. Chances are that someone searching for herms or dickgirls would be equally satisfied with this picture because they're looking for a subject. You could make a system where this might be tagged as "penis breasts mouse implied[dickgirl anal_penetration]" and when searching there could be a check box for "include implied tags".
Updated by anonymous
Conker said:
Well what did you think the strapon was modeled after? A toaster oven? :V
It was modeled after a penis. Which is not exclusively male on this site.
There are many definitions of male. Since I've worked as an entomologist for a couple of decades, the biological one is ingrained in my brain:
Organism that produces sperm cells is male, whereas one that produces egg cells is female; one that produces both is a hermaphrodite. The appearance of the genitals and the secondary sex characteristics have no actual bearing in determining the gender.
But obviously, that definition would be completely impossible to adapt for TWYS. So I've been trying to adapt to the idea that the gender is somehow universally determined by breasts or vaginas. But it still seems quite alien to me.
Updated by anonymous
Aurali said:
well.. it would give a girl a penis.. so a dickgirl .-.
Na :V that would just be tagged as a female with a strapon :D
Updated by anonymous
Conker said:
Na :V that would just be tagged as a female with a strapon :D
but what if you can't see the straps! :O
Updated by anonymous
Aurali said:
but what if you can't see the straps! :O
strapless strap-on (United States Patent Number 5690603)
Updated by anonymous
Edit: I have no idea why this ended up here from the RBUR panel. I'll see about getting it removed