[X] My Assistant

When searching for a manga or doujinshi you will find pages and pages of different translations and versions (digital, uncensored rewrites etc). So I figured it would be a good thing collect all these under one "page".

And because i suck at explaining through text, here's an example picture:

(IMG:[i2.lulzimg.com] http://i2.lulzimg.com/97908f62f2.png)


This would give a much batter overview whats available, it could also gather up all tags so they match on the different galleries.

Thoughts? (IMG:[invalid] style_emoticons/default/smile.gif)



This post has been edited by binglo: Mar 31 2012, 20:26

- Someone would have to do the linking manually (and keep everything updated). With over 100,000 galleries in the manga and doujin sections alone, that could be... problematic.
- Some of the more popular galleries just have too many variations and the whole thing would become a mess:

Raw | Raw (Uncensored) | Raw (Digital) | English (LWB) | English (Yoroshii) | English (LWB) (Uncensored) | English (Rewrite) | Chinese | Chinese (Uncensored) | Korean | French | Spanish | Portuguese (Uncensored) | Russian

- Some galleries are collections of multiple chapters/works; having to link multiple galleries could make things messy (especially when coupled with the previous point)

- Someone would have to do the linking manually (and keep everything updated). With over 100,000 galleries in the manga and doujin sections alone, that could be... problematic.

Indeed. Tho the system would ever so slowly expand, the immediate result would be on stuff thats released "now".
The "old" system would still be valid, and could be added gradually. Question is how (or by who) the linking would be done - something I've yet go figure out. (IMG:[invalid] style_emoticons/default/unsure.gif)

- Some of the more popular galleries just have too many variations and the whole thing would become a mess:

Raw | Raw (Uncensored) | Raw (Digital) | English (LWB) | English (Yoroshii) | English (LWB) (Uncensored) | English (Rewrite) | Chinese | Chinese (Uncensored) | Korean | French | Spanish | Portuguese (Uncensored) | Russian

This could be avoided by have language groupings:

+ Raw | + English | + Chinese | Korean | French | Spanish | Portuguese (Uncensored) | Russian

Pressing anything with a "+" would expand that table. Also of course have a scrollbar, when it gets too long.

- Some galleries are collections of multiple chapters/works; having to link multiple galleries could make things messy (especially when coupled with the previous point)

My idea was only concerning stuff thats is the same material, so a single chapter would not get linked to a full volume of vise versa, maybe a little bar to where the full "work" can be found should be added? (IMG:[invalid] style_emoticons/default/smile.gif)

This post has been edited by binglo: Mar 30 2012, 17:14

As an idea, I think it is good. And about the need of doing it manually, well, galleries are at the moment tagged manually.

So, if it gets done, I doubt every gallery would ever get linked correctly, but at least the popular ones would, and that is a serious improvement in my opinion.

And there could be extra encouragement to link the galleries to (correct) original/translated groups by giving extra glitterpoints for doing so.
It should be quite obvious which manga needs to be (translation)grouped under where what, by looking at the images.

When searching for a manga or doujinshi you will find pages and pages of different translations and versions (digital, uncensored rewrites etc). So I figured it would be a good thing collect all these under one "page".

And because i suck at explaining through text, here's an example picture:



This would give a much batter overview whats available, it could also gather up all tags so they match on the different galleries.

Thoughts? (IMG:[invalid] style_emoticons/default/smile.gif)

What do you do when there are two of the same type and no one can tell who one or both translators are? Two different Korean translations for example. Even if I could identify them, I could not write their names. You would end up with ....

korean | korean or 한국어 | 한국어

Also, to be really effective the language's "name" would have to be in the native language.

Gallery grouping is a fine idea, but what you've presented only seems to help once you're viewing a gallery. You haven't really described how the display of these galleries is condensed in the search results. That actually has a lot more potential problems.

I do like the basic idea implied by your mock-up, though. I'm just not sure how much it's worth if the search results still list every single version separately. I mean, if you're going to pick one gallery to open from the list of alternate versions in the search results, it's probably going to be the one you want (most) anyway, so listing the other versions in that gallery will be of limited usefulness.

What do you do when there are two of the same type and no one can tell who one or both translators are? Two different Korean translations for example. Even if I could identify them, I could not write their names. You would end up with ....

korean | korean or 한국어 | 한국어

The gallery names either will or will not make a distinction, themselves, so simply follow that. If you're going to list them out compactly and the gallery titles of the two translations in the same language make no explicit distinction, then call it something like "Korean A" and "Korean B". If you're going to list them out longer (I'm actually a little partial to the alternate galleries appearing in a pop-up like favorites or archive download), then just use the gallery name in full, as is.

Also, to be really effective the language's "name" would have to be in the native language.

Alternatively, you could take a page from anidb's book and use flags. Or some combination.

This post has been edited by PeopleDon'tDanceNoMore: Mar 31 2012, 08:40

Sorry for my inability to explain what I'm thinking with words, pictures to answer your questions (IMG:[invalid] style_emoticons/default/smile.gif)

What do you do when there are two of the same type and no one can tell who one or both translators are? Two different Korean translations for example. Even if I could identify them, I could not write their names. You would end up with ....

korean | korean or 한국어 | 한국어

Also, to be really effective the language's "name" would have to be in the native language.

(IMG:[i2.lulzimg.com] http://i2.lulzimg.com/3820ea815d.png)

When pressing the language of your choice the alternatives underneath the black arrows would appear.

Gallery grouping is a fine idea, but what you've presented only seems to help once you're viewing a gallery. You haven't really described how the display of these galleries is condensed in the search results. That actually has a lot more potential problems.

I do like the basic idea implied by your mock-up, though. I'm just not sure how much it's worth if the search results still list every single version separately. I mean, if you're going to pick one gallery to open from the list of alternate versions in the search results, it's probably going to be the one you want (most) anyway, so listing the other versions in that gallery will be of limited usefulness.

When just browsing, this is how it looks now:

(IMG:[i2.lulzimg.com] http://i2.lulzimg.com/5166f3d26b.png)


How it would look with this system (for a person that has selected "English" as preferred language):

(IMG:[i2.lulzimg.com] http://i2.lulzimg.com/99fb264dd0.png)


Only one cover is shown for each "unique" material, effectively cutting down multiple pages on big mangakas.
To further simplify the whole language issue, user could specify what language they are speaking (checkboxes under settings). So when they later browse, everything that's translated to their language shows up as such. Stuff missing in preferred language would default to raw.

How i imagine it: People searching for an mangaka, for example Kisaragi Gunma, it would still only show one "cover" per material, clicking on his choice would bring the user directly to the language he was searching for. If there is something the user not happy with he can then easily navigate to all other releases of that material from the "on page menu".

Also as i said before, this would make tagging easier and much more consistent on multi release stuff, only tags that would be unique per release would be: language, translated/rewrite and uncensored (did i forget any?)

This post has been edited by binglo: Mar 31 2012, 20:28

Gah, I don't like thumbnail mode =/ , or whatever that's called...but that's me, and irrelevant.

This pretty much only works for Doujin/Manga, so any searches including other categories probably couldn't have this feature...which probably means you'd have to access this search in a special way.

Then you need to be able to set a language. And the system needs to filter all the search results for that (or raw...or the earliest posted version if neither is available). That seems like a lot of work for the system, to me...I'd love it, I just don't know that it's the best solution. On the other hand, if it's a special search, perhaps it could be performed in a way that's not too taxing.

Being able to set a language, they must also be defined. Sure, we have language tags right now, but the system couldn't pick out which tags are languages and which are everything else. So the languages have to be partitioned off and listed formally. But that's not a big problem, really.

Then I start wondering whether individual galleries would be designated as English/Japanese/Chinese/etc, or whether that language information would only attach to the links between versions (for a work W, there's entries E_i, each entry having data V_i (version i: the address of the gallery) and L_i (the language of version i)...and N_i probably (name of this link; e.g., "A", which combines with L_i to form "Korean A")).

And then how do you know what's raw? Doujin/Manga are not exclusively Japanese...I think? This may have been a greater point for some non-h commercial manga that are originally Korean or Chinese (which are now wiped), but I have no clue if there's actually any Doujin/Manga in e-h that aren't originally Japanese...


...There's probably other points, but I've had enough brainstorming for now. Other people can raise more issues. Although, really, I'm not sure whether what I've stated are "big" problems or just questions of the implementation which need to be answered (and possibly could be, in an adequate manner for all I know)...food for thought, in any case.

Tenboro's the one who really knows how it all works and can therefore say how feasible something might be. In the meantime, we can argue among ourselves...

Gah, I don't like thumbnail mode =/ , or whatever that's called...but that's me, and irrelevant.

This pretty much only works for Doujin/Manga, so any searches including other categories probably couldn't have this feature...which probably means you'd have to access this search in a special way.

Indeed the idea was spawned from the need I saw with manga/doujinshi, but the system could also work on some special cases for artist/game cg's. Often very big collections spawn multiple or at least a couple of galleries, these could easily be linked. Also uploaders that separates cg from character cutouts, environments and such could cross link stuff:

Game CG Part 1 | Game CG Part 2 | Character cutouts

Then you need to be able to set a language. And the system needs to filter all the search results for that (or raw...or the earliest posted version if neither is available). That seems like a lot of work for the system, to me...I'd love it, I just don't know that it's the best solution. On the other hand, if it's a special search, perhaps it could be performed in a way that's not too taxing.

Being able to set a language, they must also be defined. Sure, we have language tags right now, but the system couldn't pick out which tags are languages and which are everything else. So the languages have to be partitioned off and listed formally. But that's not a big problem, really.

The system is not supposed to change the now working structure - instead build upon it. So a person clicks in preferred language in settings, after that the only thing the search does is a normal "tag search" for what language was picked. In other words, it should not be more taxing then if a person just did a search in todays system.

Then I start wondering whether individual galleries would be designated as English/Japanese/Chinese/etc, or whether that language information would only attach to the links between versions (for a work W, there's entries E_i, each entry having data V_i (version i: the address of the gallery) and L_i (the language of version i)...and N_i probably (name of this link; e.g., "A", which combines with L_i to form "Korean A")).

I don't understand this (IMG:[invalid] style_emoticons/default/unsure.gif)

And then how do you know what's raw? Doujin/Manga are not exclusively Japanese...I think? This may have been a greater point for some non-h commercial manga that are originally Korean or Chinese (which are now wiped), but I have no clue if there's actually any Doujin/Manga in e-h that aren't originally Japanese...

Raw = Original language, no matter what language that is.
For example a western magazine thats original in English get translated to Spanish. A user has his preferred language set as French, does a search for said magazine, the system only finds a Spanish and English version, system defaults to English.
Seeing as everything thats non-raw gets either the translated or rewrite tag it should not be to hard to link that.

Tenboro's the one who really knows how it all works and can therefore say how feasible something might be. In the meantime, we can argue among ourselves...

Indeed Tenboro is the one thats holding all the cards, still if the community shows interest it would seem a lot more relevant to integrate. (IMG:[invalid] style_emoticons/default/smile.gif)

This post has been edited by binglo: Mar 31 2012, 15:58

The system is not supposed to change the now working structure - instead build upon it. So a person clicks in preferred language in settings, after that the only thing the search does is a normal "tag search" for what language was picked. In other words, it should not be more taxing then if a person just did a search in todays system.

Right now, the system has NO way to show, in a search for Gunma, work X's English version (if the galleries have English and raw versions for X) and also gallery Y's raw version (if the raw is the only version of Y), not in the same search. If you searched Gunma English, you wouldn't get Y raw; if you searched Gunma -translated, you wouldn't get X in English. So it's not quite a "normal" tag search.

First the system has to get the raws for your search. Then it has to get the galleries in your target language. THEN, for each translated gallery it found, it has to check whether it was also represented in the raw set in order to prevent that raw gallery from showing up in the final results. (And then there's the question of galleries that are translated, but not in your target language and that don't have a raw...) So no, a "normal" approach really isn't that simple. In fact, it seems to have quite terrible performance.

All this isn't touching how finding the "raw" version in a "normal" search pretty much means saying -translated, and if you don't have any other search terms, that's a negative-only search, which Tenboro has been against. So unless you can only use this search with another term, your "normal search+" approach is already sunk.

Raw = Original language, no matter what language that is.
For example a western magazine thats original in English get translated to Spanish. A user has his preferred language set as French, does a search for said magazine, the system only finds a Spanish and English version, system defaults to English.
Seeing as everything thats non-raw gets either the translated or rewrite tag it should not be to hard to link that.

First: linking isn't the problem, searching is.

And honestly, half my problem is that, if you can only identify it by process of elimination, that's an ugly "solution". The other half, though, it that this process of elimination, again, means you may have to call a negative-only search, which, again, would leave your idea sunk before it even started.

Even if you had a positive term, though (instead of relying on process of elimination), that doesn't alter the process I mentioned above (i.e., the one with terrible performance). Maybe there's a better way, but I can't imagine what it would look like if you're trying to not change the system.


I didn't think about it too hard before, but I really don't think you can hope to dump this directly on top of the current system with no changes besides linking the galleries. It may be possible if you specially search the linked groups in some fashion, but a "normal tag search" or anything like it will NOT work.


Edit: Oh! Another problem...

Suppose there's a doujin (in Japanese, raw) called "Uma" (馬, horse). It's been translated into English as "The Horse" and into Spanish as "El Caballo". Suppose a Spanish speaker only knows it by its English name. If he searches for "The Horse" (quoted), he should get the result for "El Caballo", right?

1) This is disorienting. (This is more a problem in general.)

2) How do you make that jump? If it's just a normal tag search, "The Horse" Spanish is never going to get you the gallery for "El Caballo". So again, a "normal tag search" alone isn't going to cut it.

This post has been edited by PeopleDon'tDanceNoMore: Mar 31 2012, 18:15

Right now, the system has NO way to show, in a search for Gunma, work X's English version (if the galleries have English and raw versions for X) and also gallery Y's raw version (if the raw is the only version of Y), not in the same search. If you searched Gunma English, you wouldn't get Y raw; if you searched Gunma -translated, you wouldn't get X in English. So it's not quite a "normal" tag search.

First the system has to get the raws for your search. Then it has to get the galleries in your target language. THEN, for each translated gallery it found, it has to check whether it was also represented in the raw set in order to prevent that raw gallery from showing up in the final results. (And then there's the question of galleries that are translated, but not in your target language and that don't have a raw...) So no, a "normal" approach really isn't that simple. In fact, it seems to have quite terrible performance.

All this isn't touching how finding the "raw" version in a "normal" search pretty much means saying -translated, and if you don't have any other search terms, that's a negative-only search, which Tenboro has been against. So unless you can only use this search with another term, your "normal search+" approach is already sunk.

First: linking isn't the problem, searching is.

And honestly, half my problem is that, if you can only identify it by process of elimination, that's an ugly "solution". The other half, though, it that this process of elimination, again, means you may have to call a negative-only search, which, again, would leave your idea sunk before it even started.

Even if you had a positive term, though (instead of relying on process of elimination), that doesn't alter the process I mentioned above (i.e., the one with terrible performance). Maybe there's a better way, but I can't imagine what it would look like if you're trying to not change the system.
I didn't think about it too hard before, but I really don't think you can hope to dump this directly on top of the current system with no changes besides linking the galleries. It may be possible if you specially search the linked groups in some fashion, but a "normal tag search" or anything like it will NOT work.

Very true (IMG:[invalid] style_emoticons/default/smile.gif)
I guess, if its too taxing on the system it would just work as it does now, if nothing is found nothing turns up.
The concept would still stand: gathering all different versions of each material under one "cover", making it a lot easier to survey what is available.

Edit: Oh! Another problem...

Suppose there's a doujin (in Japanese, raw) called "Uma" (馬, horse). It's been translated into English as "The Horse" and into Spanish as "El Caballo". Suppose a Spanish speaker only knows it by its English name. If he searches for "The Horse" (quoted), he should get the result for "El Caballo", right?

1) This is disorienting. (This is more a problem in general.)

2) How do you make that jump? If it's just a normal tag search, "The Horse" Spanish is never going to get you the gallery for "El Caballo". So again, a "normal tag search" alone isn't going to cut it.

Good point, I guess this would need some fixing. Just as all the tags should be shared between different versions, why not names?

On another note, I'm far far knowledgeable how the actual system works, more had a good concept image in my head (IMG:[invalid] style_emoticons/default/smile.gif)

This post has been edited by binglo: Mar 31 2012, 18:55

Myself I just wonder these cases, when there is only translated version available, like only Chinese version. I prefer English, Finnish and Japanese versions and excluding everything else out of search, but that Chinese would be fine then if there isn't anything else available.

So some "Show only only following languages if galley isn't first upload on group"

Good point, I guess this would need some fixing. Just as all the tags should be shared between different versions, why not names?

That sounds like a great idea, but no, don't share tags, at least not as the system currently is. First off, languages are identified by tags right now, and those can't mix. Similarly for decensored and color (if color is added). Also, what if one of the galleries is a rewrite? I'm not sure about this, but since they make up the story, is it possible they've got a few tags the raw/translated versions might not have? (Especially, I think they add incest a lot.)

I do think it would be great if new galleries that are a new version of previous galleries could start off with a fundamental set of tags from previous versions, but THAT would definitely require some changes...

Myself I just wonder these cases, when there is only translated version available, like only Chinese version. I prefer English, Finnish and Japanese versions and excluding everything else out of search, but that Chinese would be fine then if there isn't anything else available.

So some "Show only only following languages if galley isn't first upload on group"

Giving one main option and defaulting to raw makes perfect sense and is kind of the bare bones requirement for this idea. Giving any more options past the first quickly adds to your performance cost and starts making less sense.

This is a good reason for why I think maybe this sort of search should somehow show all the languages an individual search result has on the search page. (Something like flags at the beginning of the line?)

This also raises the question, though, of how to prioritize different versions of a translation (if at all, or else how to distinguish them absolutely minimally (space being at a premium in the search results)). This is also a valid question if it's your target language: if SaHa and LWB versions are out there, which English version gets priority? Or how do I make that distinction fairly, i.e., so that users aren't going to miss the fact that there's two versions and will choose the one they really want (if they would have a preference) rather than whichever link is more prominent?

Good point, I guess this would need some fixing. Just as all the tags should be shared between different versions, why not names?

Gallery Names should be the same for all languages, making allowances for localized name. However, many Euro language galleries use the English name (or translation thereof) as most are translated from an English version. (They need to be eventually renamed to the Japanese Title.)

Also searching for Kisaragi Gunma +English would produce the same results as your second image and reveal the translator and censorship status (if relevant). Plus, it eliminates the English title issue.

That sounds like a great idea, but no, don't share tags, at least not as the system currently is. First off, languages are identified by tags right now, and those can't mix. Similarly for decensored and color (if color is added). Also, what if one of the galleries is a rewrite? I'm not sure about this, but since they make up the story, is it possible they've got a few tags the raw/translated versions might not have? (Especially, I think they add incest a lot.)

I do think it would be great if new galleries that are a new version of previous galleries could start off with a fundamental set of tags from previous versions, but THAT would definitely require some changes...

I guess a lot of suggestions that came after the initial idea should be shaved off, if this should ever be possible to implement. (IMG:[invalid] style_emoticons/default/smile.gif)

So if we go back to the basic idea:

Having a link bar inside galleries, for easy navigation between them and overview:

(IMG:[i2.lulzimg.com] http://i2.lulzimg.com/3820ea815d.png)

Gallery Names should be the same for all languages, making allowances for localized name. However, many Euro language galleries use the English name (or translation thereof) as most are translated from an English version. (They need to be eventually renamed to the Japanese Title.)

Also searching for Kisaragi Gunma +English would produce the same results as your second image and reveal the translator and censorship status (if relevant). Plus, it eliminates the English title issue.

I agree that the system is working fine as it is now if you actually have a name and already know a specific work your interested in.

The way I see it: people login to https://e-hentai.org/, they see a new or popular release, lets say in English. They click on it and sees the "navigation bar", showing that their preferred language is available.
The strong point of the suggestion would be in helping people that browse/skims through galleries either by only typing in a mangaka (to see everything thats available) or even just looking through 1-5 pages of latest released stuff.