Topic: Should we provide a 'training area' to all members so that they can test their abilities/understanding on tagging?

Posted under General

Hello. Apologies if this belongs in off-topic or feature requests, I think this falls more into the "general bucket", but please feel free to relocate this if appropriate (I considered that this is perhaps a feature request, but I think it's more of a tool, but again, please relocate if necessary).

Anyway, before continuing, I'd like to provide some brief context. Over the past few weeks, I've been on E621 quite a lot (probably more than I should), though it is only the past week or so that I've started to become somewhat active in the community, by tagging, updating pools, and creating discussions (posts?) such as these.

During that time, I've tried my best to tag appropriately, though at times this was not always achieved. Along the way, the community kindly provided me feedback on where I want wrong and what I should have done right, to which I greatly appreciate. Anyway, the reason why I bring this all up is to lead into the reason for creating this discussion, in what I would like to propose a sort of training area for all E621 members.

Such a training area could be something akin to what you find on visual Capacha's (you know, like "find all the trains", "find all the roadwalks", "find all the imposters", etc etc), a multiple choice quiz, or a yes/no style quiz, or perhaps a mix of all of the above.

The first type (visual captchas) may be useful for training people on differing species. Everyone can tell the difference between an equid and a scalie, but sometimes, especially with anthro, I find it's hard to distinguish between felines and canines, or mustelids and felines.

The second type (multiple choice quizzes) may be useful for approximating ones general tagging abilities. I don't think there's anyone out there who can claim that every tag on every post they've edited was correct, there's bound to be some human error, but such a multiple-choice style quiz could at least approximate ones general tagging ability, to which there could be automatic feedback on what went wrong, and, most importantly, why.

The third type could be used to specifically home in on tags which cause the most confusion for the majority of users. For example, new users may not be up to scratch on the TWYS protocol, and I speak from personal experience here. I had at one point tagged a charachter within a comic who was canonically equid by species, to which it was also clear when looking at their facial features, as being just that. This is all well and good when you could see their face, but the problem is that, in some posts at least, you could only see their chest, to which anyone stumbling upon that post in and of itself would say that they were clearly human, and yet still I had tagged them as equid. This intuitively made sense to me given what I knew outside of that post, but now that I understand the TWYS protocol, and the reasoning behind that protocol, it's clear that the human tag was the appropriate choice.

To me, the TWYS protocol is probably one of the most important, if not thee most important, protocol on E621, given that it applies to every single post. Thus, in order for users to test their understanding of this protocol, they could be shown a series of images, either from the same comic, a similar one, or perhaps even a series of original images created for the test itself. Each image after the next could provide less and less background information, but to which ones understanding of it simutaneously becomes more and more dependent upon the information provided in the images preceding it. For example, you could display a minotaur (head, tail, and hooves of a bull, with the body of a man) which at first showed a full body, but with each image gradually honing in on specifc body parts of the minotaur. So, let's say that the fifth image displayed only the head of the minotaur. Intuitevly, we all know this isn't just a bovid chararchter, at least canonically, yet strict E621 policy via the TWYS protocol would dictate not to tag it as such, but only as a bull, and hence bovid via implication (albiet it may not look like an average bull, perhaps it looks "demon" like, so it would also have the demon tag and such, but hopefully you get the idea). It would probably be a surprise to no one if a new user tagged such a post as being a minotaur, but if so, their answer could of course be marked as wrong, but to which they could then be provided the link to the TWYS protocol, thus hopefully significantly decreasing the error rate when tagging actual posts.

By the way, I should note that I don’t intend for the above to be part of some sort of entrance exam, i.e, something that users would have to be pass in order to aquire membership, since this I believe would do more harm than good by putting up a barrier between those who could otherwise be very valuable members of the community. However, I think this could be made available to all members, past, future, and present, because this would, I hope, decrease the error rate for those that use it, and help better ones understanding on any recurring tag-related issues one may be facing.

I am just a user/consumer of the site, and have nothing to say about this site policies.

I just want to bump this thread up, because I think your writting deserves some consideration. Regarding just part of what you say I am myself surprised at the correction in which the tagging of animals tries to follow correct principles of scientific zoological taxonomic nomenclature. Maybe not at a "expert level", but still seem to me not ready available to just anyone that have not read a bit about it.

Regarding other tagging aspects, I don't even think I could tag "well" any image that I could eventually uplodad. When I tried to upload some images of a comic (that was deleted), I was surprised at the level of intelligence already built up in the system, to direct such tagging. I wish that I could have had the oportunity to learn more about it.

Still, I just bump once your proposal, that altough is not so much relevant for me now, looks interesting, tough.

  • 1