Donmai

It would be a lot easier to hunt down missing tags with three tags

Posted under Tags

I know that's a premium feature, just wanted to vent my frustration. The last couple days I wanted to hunt down the missing no_shoes tags, and filtering it out from single tags is... daunting to say the least.
I split the workload by cutting it in half with the favcount, but still damn it's a lot to go through.

I have no idea what would be the ideal solution, obviously you can't just give out free tags if that's the business model. But maybe negative tags should be half-price or something, I dunno.

That would be impossible right now since we don't have a payment processor at the moment. You can learn more about the current situation in topic #21157.

We occasionally hold raffles giving away Platinum upgrades to members to help make up for the inability to pay for Gold. We just held one a few days ago in topic #27803 and we'll probably have another in a few months. If you don't think you'll be able to help out enough to reach Builder, you can try entering the next raffle for a shot at winning an upgrade.

Two tags is a pretty extreme limit IMO. seems like the limit should be three or four. Adding to the above you can't even do two characters and a specific artist for instance. Or two characters and a specific act. Which seems like what most people would be looking for.

Blank_User said:

That would be impossible right now since we don't have a payment processor at the moment. You can learn more about the current situation in topic #21157.

We occasionally hold raffles giving away Platinum upgrades to members to help make up for the inability to pay for Gold. We just held one a few days ago in topic #27803 and we'll probably have another in a few months. If you don't think you'll be able to help out enough to reach Builder, you can try entering the next raffle for a shot at winning an upgrade.

Look, I'm not trying to ruffle any feathers here, it's not my website (and not yours, I know) and I've read the discussion regarding the gold accounts for the following ~half a year.
But, in the light that we conceivably might never find a payment processor (or at least not for a very, very long time) I think it makes sense to take a critical look at how the site is structured and what should remain behind the gold account paywall and what shouldn't, and what ways there should be for obtaining an account upgrade (to either platinum or builder) there should be, what the requirements are, and what behaviour we do encourage with this and which behaviour we want to encourage.

Raising the tag count for all users by 1 or introducing an intermediate rank might be solutions that keep the people engaged and, well, visiting danbooru, instead of frequenting other websites.
Another solution might be to make the raffles for platinum way more common then they currently are, like a weekly event, or something like that.
I'm not saying that any of those steps are mandatory or even the right course of action, I'm just saying that, at least to me, the argument of "we can't do it because thats only for gold accounts" starts to sound less and less like a good argument.

Again, I want to stress that I'm not trying to start a fight, and especially not with you Blank User :) I know you are just forwarding commonly cited arguments, and those are the ones I'm intending to address, not you specifically. Hope we all can have a respectful discussion about it.

Let me close with that there are reasons why I still prefer danbooru over the competitors. It's really mostly due to the largely professional and respectful tone, the good moderation, and that the user base isn't comprised out of 90% fakes and hornyposters. That are things that I personally value immensly, and thats why I want to see danbooru survive and thrive.

Just my 2 cents :)

CasualStan said:

Again, I want to stress that I'm not trying to start a fight, and especially not with you Blank User :) I know you are just forwarding commonly cited arguments, and those are the ones I'm intending to address, not you specifically. Hope we all can have a respectful discussion about it.

No worries. I'm not saying this is necessarily the best way to run the site. I only bring up these points to keep other users informed about the current situation and their options.

One thing I forgot to mention is that the blacklist can be used as a workaround for limits on negative tags. They're still show up on the page (though hidden) so it's not a perfect substitute, but you can put in as many as you want. I used it to help with tag gardening while I was still a basic member.

Blank_User said:

No worries. I'm not saying this is necessarily the best way to run the site. I only bring up these points to keep other users informed about the current situation and their options.

One thing I forgot to mention is that the blacklist can be used as a workaround for limits on negative tags. They're still show up on the page (though hidden) so it's not a perfect substitute, but you can put in as many as you want. I used it to help with tag gardening while I was still a basic member.

I see what you mean :) great tip regarding using the blacklist to circumvent the tag limit :)

CasualStan said:

[...]
But, in the light that we conceivably might never find a payment processor (or at least not for a very, very long time) I think it makes sense to take a critical look at how the site is structured and what should remain behind the gold account paywall and what shouldn't
[...]

Raising the tag count for all users by 1 or introducing an intermediate rank might be solutions that keep the people engaged and, well, visiting danbooru, instead of frequenting other websites.

That is more or less the issue from what I understand:
At the moment there is no income. No income means that the website is held up exclusively by the remaining subscription money gathered throughout the past years and evazion's own pockets.
Raising the amount of tags for all user would lead to an increased server usage, so higher costs.
Danbooru is probably not going down anytime soon, but I can see why certain things haven't been made available.
(On a side note - I completely agree with you and have been a staunch supporter for at least 3 tags for all users since the beginning, but I can see why it has not been done these days.)

Login_to_view said:

That is more or less the issue from what I understand:
At the moment there is no income. No income means that the website is held up exclusively by the remaining subscription money gathered throughout the past years and evazion's own pockets.
Raising the amount of tags for all user would lead to an increased server usage, so higher costs.
Danbooru is probably not going down anytime soon, but I can see why certain things haven't been made available.
(On a side note - I completely agree with you and have been a staunch supporter for at least 3 tags for all users since the beginning, but I can see why it has not been done these days.)

Thats a fair take and something I didn't consider, damn.
But maybe this could be circumvented with another user tier which is just more freely given to people who are contributing period, maybe even semi-automatically? The details would of course still be needed to be hashed out, but something like "all users that upload 100 aproved pictures/make 100 edits qualify for the casual rank", which are then manually approved to avoid bots from farming those, or something like that. I'm sure there are people smarter than me that could figure something out.

Heartcatcher said:

TBH I agree with the post but I also feel like ideally metatags just shouldn't be taken into account... Especially order:*, it's silly I can't search something like earphones -digital_media_player order:id.

Sounds like a good compromise to me Heartcatcher :)

Order searches can get really expensive database-wise, depending on which order you're sorting by. For example, order:id is almost instant because the ID column of the posts table is indexed, but order:comment_count has to join the posts table with the comments table to figure out which post has the most comments, which is much much slower because it's not a precomputed information. In fact that last search times out even for me.

nonamethanks said:

Order searches can get really expensive database-wise, depending on which order you're sorting by. For example, order:id is almost instant because the ID column of the posts table is indexed, but order:comment_count has to join the posts table with the comments table to figure out which post has the most comments, which is much much slower because it's not a precomputed information. In fact that last search times out even for me.

Ah, that's unfortunate. I do wish at least order:id would work, but I doubt that's possible behind the scenes?

1