<-- Archived --> *Image tagging concept

[Xela]

Quick note to Xela: I still have no idea why you think that a hierarchical system doesn't work - you see problems where there are none. If working in a casino is part of the maid job, the maid outfit can be used just as well. And if the tag changes in order to use a fallback, so can the text description.

Quick note to living:

Would you mind writing that one text?
============================================
By your design:

If we have a parent tag: MaidJob, we do the action.

Action = Dealer. (You don't like dealer tag or going to say it's a different parent category? Lets say: Serving tables.)

No pic was found for action... so we fall back to MaidJob. Action is still Serving Tables.

In MaidJob category we have: Pics of a girl cooking a dinner, pics of a girl cleaning the house with a feather duster...

We failed to find the exact tag... so we do the action (corresponding stat/monetary rewards/clients satisfaction modification) and we fall back on a general text describing Serving Tables action while showing either a cooking dinner or a cleaning picture. So, write me that text...

I've never seen a well working system built the way you do it before, I'm curious how you'll see this in a year or two.

As it stands now, I can promise you that you will (and propably already did) lose a lot of potential pack creators, simply because your system is so "extensive" that nobody wants to deal with it (except for very dedicated people).

Also, I wish you lots of fun while doing exactely what you wanted to avoid with your system: Retagging. It propably won't be the last time.

Our system is working marvelously. Pictures are very well matched with their jobs, fallbacks are developer defined and extendible without bothering content creators.

You may not have noticed but we didn't even write a tagging guide yet, because we always knew tagging system is unfinished and will require one last review before we feel justified asking modders to contribute.

No, it will be the last time.

[Pinkutako]

No, no.  What you guys are missing is what he has wanted to hear all along.   He wants this:  "Oh, wow, Livingforever, you are the smartest, bestest programmer of all time.  How can we not worship your every thought.  Let us completely drop our concept that we have been working on for over a year and do things your way, because you are Soooooo smart."


Ya,  nothing like a someone trying to railroad their own, unsolicited ideas into a project and then get huffy when no one listens.




From my end user viewpoint.  Livingforever is missing the entire point of the tags system.  Which is making the ingame text not only match the image, but have the ability to describe the image in detail.  This is the defining trait of this game IMHO. 

[Xela]

No, no.  What you guys are missing is what he has wanted to hear all along.   He wants this:  "Oh, wow, Livingforever, you are the smartest, bestest programmer of all time.  How can we not worship your every thought.  Let us completely drop our concept that we have been working on for over a year and do things your way, because you are Soooooo smart."


Ya,  nothing like a someone trying to railroad their own, unsolicited ideas into a project and then get huffy when no one listens.

Clearly an overkill...

He's basically trying to do what (imo) cannot be done, it's not like we haven't entertained the idea of hierarchical tags before. Noone could make it work nearly as well as a system where sorting/image picking logic is done in code on per case basis.

From my end user viewpoint.  Livingforever is missing the entire point of the tags system.  Which is making the ingame text not only match the image, but have the ability to describe the image in detail.  This is the defining trait of this game IMHO.

**Ability to match picture to the setting/surroundings/atmosphere.

With the next release, we'll document the event system and tagging. Events are practically written in open text at this point.

g "I am a bit hungry!"
g "Want to go out and eat?"

is all it takes. Showing/transforming pictures is also very easy. With out tagging concept, events can be written with easy without any particular girl in mind while preserving very accurate image representation and specify fallbacks as required.

[DarkTl]

So, no new tagger after all, I take it? Too bad.

[Xela]

So, no new tagger after all, I take it? Too bad.

? I am not sure I got that from his post...

Want me to code one to PyTFall? I could be done with a base prototype by tonight? (75% of logic required is already there).

[DarkTl]

Would be cool. His last post sounds like an upset goodbye to me.

Though I'm not sure if it should be binary or name based, it's up to you, I guess.
And since we have a permanent coder for this one, it doesn't have to be purely xml/json based if it will be too difficult to implement.

I'll remind you what Rudi didn't finish in his tagger:
- an option to delete currently displayed picture; moreover, with his tagger you cannot even delete any picture in the folder, it just stops working;
- jump to N option or something like that; I personally like how mcmanus solved this problem.

[Xela]

Would be cool. His last post sounds like an upset goodbye to me.

Though I'm not sure if it should be binary or name based, it's up to you, I guess.
And since we have a permanent coder for this one, it doesn't have to be purely xml/json based if it will be too difficult to implement.

I'll remind you what Rudi didn't finish in his tagger:
- an option to delete currently displayed picture; moreover, with his tagger you cannot even delete any picture in the folder, it just stops working;
- jump to N option or something like that; I personally like how mcmanus solved this problem.

Ok, I cannot do a layout like Rudi's where you can move stuff around. Or rather I can but it'll take a good deal of time (don't think I can manage that by tonight). I am not going to bother with XMP either, there is little point in that.

I'll just come up with the most basic version. We'll all test it and add stuff to it as required.

PS: It's not really up to me... I need opinions:

Binary:
+ As fast as it's going to get
- Not readable without software
- Cannot add/get rid of an image you dislike without software

Files:
- Somewhat slower than binary
- There could be a very unlikely case where image name will get too long (Very unlikely)
+ Humanly readable
+ Can just delete any image you don't like without having to adjust anything anywhere
+ Can add an image without tagger just by memorizing basic tags (if we go with a pretty format obviously)

[DarkTl]

Hm, yeah, I forgot about deleting... Well, we do want players to delete any pictures that they don't like without too much fuss. On the other hand, the tagger might be a part of the release, so anyone could do it in the end.

How difficult it might be to delete a picture with binary type? I mean, could it be done by pressing one button after deleting all unneeded pictures, like "rebuild pack" or something? Or you have to manually delete all deleted pictures again?

[Xela]

How difficult it might be to delete a picture with binary type? I mean, could it be done by pressing one button after deleting all unneeded pictures, like "rebuild pack" or something? Or you have to manually delete all deleted pictures again?

Difficult to say. It should be doable: writing a script that forces a check of existing images vs database entries and deletes entries as required over all the images in the game... to be executed as requested by user.
Edit: This is doable but a bit clumsy due to a database structure....

On the other hand, filenames are just loaded at the game start or when a database rebuild is requested.

[DarkTl]

Ok, what about names approach? Does the game check all existing pictures every launch? Or we'll need an option to rebuild database after deleting pictures anyway?
Because right now with jsons approach you do need to run tagger and rebuild json after deleting, and it's not a big deal. Even if we wouldn't change our approach, players still couldn't delete pictures 100% freely.

[Xela]

Does the game check all existing pictures every launch?

You delete the picture and that's it. Same when editing a tag inside of a filename manually. Every time new game is started, database will be build based off the filenames. Database inside of the game will always have to be rebuild, doesn't matter what approach we take. Otherwise we'll have to create the database every time a save is loaded instead of getting one from the save.

[DarkTl]

I see. Well, let's take players friendly approach and use names then.

[Xela]

I see. Well, let's take players friendly approach and use names then.

I agree + going to binary or even loading binary instead of filename if binary file is available is always an option.

I am killing off jsons if there is going to be an update tonight. All images will have to be converted to filenames.

PS: You/CW can help out. There is a new tagdict.rpy (or something like that) file that maps two letter combinations to actual tags. See if you want to improve that. Like some obvious tags like Profile (pr) or BlowJob (bj) don't have to be obfuscated with letter + numeral as I did it.

Once that's done, we can all rename the files and get rid of JSONs.

[DarkTl]

Done, pushed. Base tags have recognizable designations now. Impossible to do for all tags of course, at least with 2 letters.

[Xela]

Done, pushed. Base tags have recognizable designations now. Impossible to do for all tags of course, at least with 2 letters.

Kewl

We'll need tabs I think, Ren'Py is having a bit of trouble with that many textbuttons. But I haven't tried using normal buttons yet, could be a lot faster...