Credits

[jotaroraido]

So I've been speaking with a few fellows over at VGMdb, and a few concerns with the way credits were done at Moby came up. I know a few of these things have already been discussed, but I wanted to collect this all in one place. A quick braindump, if you will.

1. Sources. Obviously this has been a huge problem at Moby since day one. What's the point of doing all this exhaustive research when the sources you use to back it up get locked away in the administrator's vault for all eternity? It should be possible to add publicly-viewable sources as attachments to any data, be it images (screenshots and scans), videos, other types of attachments like text (readmes), or even simple text and links. This one's obvious, and goes hand-in-hand with...

2. Talk pages. Every developer should have their own talk page (or forum thread, as it's done on VGMdb), linked directly from their developer profile. As I mentioned before, this would make it easy to discuss a particular person's circumstances, information sources, potential merges and splits, aliases, even what the main display name should be.

3. Revisability. We're human, people mistakes. Typos, omissions, what have you. If you had an error in a credits set at Moby that wasn't a simple merge to fix, it was a huge ordeal to fix or update things, generally requiring admin intervention. You couldn't change existing headers, and role name changes had a nasty habit of getting reverted. It should be a simple matter to update existing data to fix errors or add new information. The Wiki-like free format mentioned above sounds perfect for this, so I have no doubts this will become a non-issue.

4. Japanese name readings. Yes, I'm going to keep hammering this point until it's nothing but a gaping hole in the ground. One of the biggest points of contention with the way MG did things was that a reading was required for *every* name in a translated credit set. While this was a good impetus to do research to make sure they're accurate, it unfortunately also meant that there was a lot of guesswork. One way to solve this problem would be to allow untranslated listings in credits. Another would be to have a checkbox on a developer to note that the reading of the name is unconfirmed, and perhaps highlight such names in a different color when displayed.

[Ultyzarus]

So I've been speaking with a few fellows over at VGMdb, and a few concerns with the way credits were done at Moby came up. I know a few of these things have already been discussed, but I wanted to collect this all in one place. A quick braindump, if you will.

1. Sources. Obviously this has been a huge problem at Moby since day one. What's the point of doing all this exhaustive research when the sources you use to back it up get locked away in the administrator's vault for all eternity? It should be possible to add publicly-viewable sources as attachments to any data, be it images (screenshots and scans), videos, other types of attachments like text (readmes), or even simple text and links. This one's obvious, and goes hand-in-hand with...

2. Talk pages. Every developer should have their own talk page (or forum thread, as it's done on VGMdb), linked directly from their developer profile. As I mentioned before, this would make it easy to discuss a particular person's circumstances, information sources, potential merges and splits, aliases, even what the main display name should be.

3. Revisability. We're human, people mistakes. Typos, omissions, what have you. If you had an error in a credits set at Moby that wasn't a simple merge to fix, it was a huge ordeal to fix or update things, generally requiring admin intervention. You couldn't change existing headers, and role name changes had a nasty habit of getting reverted. It should be a simple matter to update existing data to fix errors or add new information. The Wiki-like free format mentioned above sounds perfect for this, so I have no doubts this will become a non-issue.

4. Japanese name readings. Yes, I'm going to keep hammering this point until it's nothing but a gaping hole in the ground. One of the biggest points of contention with the way MG did things was that a reading was required for *every* name in a translated credit set. While this was a good impetus to do research to make sure they're accurate, it unfortunately also meant that there was a lot of guesswork. One way to solve this problem would be to allow untranslated listings in credits. Another would be to have a checkbox on a developer to note that the reading of the name is unconfirmed, and perhaps highlight such names in a different color when displayed.

1. Definitely something that is planned on being implemented. I believe one of the goals of Oregami is to be an accurate academic source, so such references are a must in all fields.

2. I second this. It was very useful for the games in MG, so why not have this possibility for all kinds of data?

3. I hope that will also be implemented, it sure seems like the thing MZ per X has already given a thought about

4. Having only the confirmed data available first, if we can edit/revise it (see item 3. ) or add more details (Romanization & Kana spelling) would be the best option to avoid mistakes...

[jotaroraido]

So it was mentioned earlier that we would most likely be using a free text format for credit listings -- obviously a strict column based approach is too limited for this sort of data -- and I've been exploring a bit on how we might handle some of the different cases. If we're going to try to keep the original formatting and text intact as much as possible, there are several things we'll need to handle:

1. Text that is not part of either a role name or a person/company/entity name.
2. Designating roles for individuals and groups of developers.
3. Designating roles for companies.
4. Multiple text sets for translated credits.
5. Multiple roles listed for a single name.
6. Probably other things I haven't thought of...?

I've constructed a quick example that shows one possibility how we might handle this.

Code: Select all

{free|キャスト|Cast}
{role|6|吹雪　ユナ|Yuna Fubuki}     {name|7325|川澄 綾子|Ayako Kawasumi}
{role|6|冴草　宗司|Sōji Saegusa}    {name|7326|櫻井 孝宏|Takahiro Sakurai}
{role|6|美咲　千歳|Chitose Misaki}  {name|7327|平松 晶子|Akiko Hiramatsu}

{role|2|Scenario}                 {name|7328|Goro}

{role|1|Planning}
{role|4|Character design}         {name|7329|Satoshi Kuramochi}

{role|3|Program}                  {name|7330|Daisuke Kawaseki}
                                  {name|7331|Kouji Saito}
                                  {name|7332|Toshio Kurisawa}
                                  {name|7333|Naomichi Sato}

{role|16|Cooperation}             {company|816|株式会社 大沢事務所|Office Osawa Co., Ltd.}
                                  {company|817|ポールトゥウィン 株式会社|Pole To Win Co., Ltd.}

Incidentally, this is taken pretty much verbatim from a real set (only cut down for space), so it works as an example of the sort of thing we'd need to handle.

Any text that is either a role name, a developer name, a company name, or a loose piece of free text that might change based on display settings, is turned into a tag. The first field says what type of text it is -- role name, person name, company, or free text -- the second gives the ID of the role number (from data lists 8 and 10) or the ID of the developer or company, the third gives the original text, and the fourth gives an optional translation/transliteration if needed or available.

Basically, a role tag indicates that any name or company tags following it will be assigned that role. Multiple successive names/companies would all apply to the most recent role tag. Multiple successive role tags would indicate that all of them (as in the "Planning / Character design" example above) would apply to the following names. The first role tag following a name would negate any previous role tags.

One thing worth noting is that both company roles and developer roles use the same tag. It's not unusual to see them credited together, so it might be ideal to combine the two role type lists (8 and 10) into a single list, or at least keep the two lists in sync with each other, and use only a generic role tag as I've done here.

When displayed, the formatting would be hidden from the user, defaulting either to the original text or the translated text. It would be possible to swap between them (or possibly even display both simultaneously), and perhaps also have an option to display the primary name of the linked developer/company instead of the given one.

Obviously, we wouldn't ask contributors to enter all these tags and such manually -- the contribution form could be set up to automate most of it. I have a few ideas on how to handle all that, but that part is still a work in progress, and I'd like to get a bit of feedback on this part before I proceed any further. I would probably write my own offline tool for formatting stuff to just paste straight in, too.

[jotaroraido]

I suppose I'll take that as an indication that it's a bit early to be thinking about this, then.

[MZ per X]

I suppose I'll take that as an indication that it's a bit early to be thinking about this, then.

Oh no, it's not too early. It's just that I'm quite busy at the moment, and can't keep up with all that's happening here. Which is good, because it means there's quite a lot happening here.

So it was mentioned earlier that we would most likely be using a free text format for credit listings

I like your approach, it's exactly what I have in mind. On a general note, we need a capable free text editor with database links (similar to the moby tags) not only for credits, but also for descriptions, reviews, walkthroughs, etc., so time spent on this is definitely not wasted. Maybe we can fork the technical basics from MediaWiki or somesuch.

6. Probably other things I haven't thought of...?

Maybe some check marks for misspelled or abbreviated names.

The first field says what type of text it is -- role name, person name, company, or free text -- the second gives the ID of the role number (from data lists 8 and 10) or the ID of the developer or company, the third gives the original text, and the fourth gives an optional translation/transliteration if needed or available.

As we're a "latin" project, wouldn't it be better to turn the third and fourth around, i. e. make the transliteration mandatory, and the original spelling optional?

One thing worth noting is that both company roles and developer roles use the same tag. It's not unusual to see them credited together, so it might be ideal to combine the two role type lists (8 and 10) into a single list, or at least keep the two lists in sync with each other, and use only a generic role tag as I've done here.

Yeah, I like that idea. Feel free to combine both lists into one in the wiki.

Overall, some good ideas. Please keep them coming.

[jotaroraido]

Cool, good to know what I'm thinking is feasible.

6. Probably other things I haven't thought of...?

Maybe some check marks for misspelled or abbreviated names.

Presumably, so they wouldn't show up on a list of aliases on the developer's rap sheet, right? Wondering if it might be more appropriate to do that in the person's listing rather than in the credits themselves...

The first field says what type of text it is -- role name, person name, company, or free text -- the second gives the ID of the role number (from data lists 8 and 10) or the ID of the developer or company, the third gives the original text, and the fourth gives an optional translation/transliteration if needed or available.

As we're a "latin" project, wouldn't it be better to turn the third and fourth around, i. e. make the transliteration mandatory, and the original spelling optional?

On the contrary, as we're an "international" project, I feel it would be best to give the original data priority, and make the translation optional. That way someone who has, say, a Polish credits set, but can't actually read Polish, can add the credits without being required to come up with a translation. Someone who does speak Polish could then update the listing with the translation at a later point. It would also mean that we could rely on the third field to *always* be the original text, and do formatting appropriately.

Also, as I've stated earlier, there's a particular problem with Japanese, as name readings can be quite ambiguous, and I would prefer to leave unknown names untranslated than make educated (or whole-cloth) guesses like we sometimes had to at Moby. They could still be linked by the kanji name, even if there's no reading given.

[MZ per X]

6. Probably other things I haven't thought of...?

Maybe some check marks for misspelled or abbreviated names.

Presumably, so they wouldn't show up on a list of aliases on the developer's rap sheet, right? Wondering if it might be more appropriate to do that in the person's listing rather than in the credits themselves...

I don't think so. If you do it in the credits, you've got the provenance for free, and you can even link to the credits where the name was misspelled, which is cool. In the person's rap sheet, we could visually separate between real aliases, and the saved misspellings / abbreviations.

As we're a "latin" project, wouldn't it be better to turn the third and fourth around, i. e. make the transliteration mandatory, and the original spelling optional?

On the contrary, as we're an "international" project, I feel it would be best to give the original data priority, and make the translation optional. That way someone who has, say, a Polish credits set, but can't actually read Polish, can add the credits without being required to come up with a translation. Someone who does speak Polish could then update the listing with the translation at a later point. It would also mean that we could rely on the third field to *always* be the original text, and do formatting appropriately.

Okay, I see your point. But I also see a language idiot like myself entering non-latin credits with a translation enclosed. For example, while I can read Russian, I wouldn't even know how to type in Russian letters. That's why I think three options should be possible: original, transliteration, or both.

[Ultyzarus]

It's true that I would also have this problem if I entered Japanese credits. Unless it is an online source where I can copy/paste the characters or if it is Kana only, I would have trouble entering any Kanji.

[jotaroraido]

I don't think so. If you do it in the credits, you've got the provenance for free, and you can even link to the credits where the name was misspelled, which is cool. In the person's rap sheet, we could visually separate between real aliases, and the saved misspellings / abbreviations.

Well, let's say a prolific developer has a commonly misspelled name. You'd have to mark it as a misspelled name in every credits set where that person shows up. If we simply left them misspelled in the credits listings and linked them all to the correct developer, we'd have a list right there where we could mark specific names as "misspelled" or "abbreviated", all on one page. There could be a display option to hide such names from their aliases, but they'd still show up with a "credited as" in the works list. Personally, I feel like keeping all information for a single person in a single place would be preferable.

Okay, I see your point. But I also see a language idiot like myself entering non-latin credits with a translation enclosed. For example, while I can read Russian, I wouldn't even know how to type in Russian letters. That's why I think three options should be possible: original, transliteration, or both.

I can see your point here. I still believe we should give priority to the original text, but at the same time, the original text for a language like Russian isn't exactly ambiguous. Perhaps there could be a flag for a credits set to denote that it's translated, but the original text is not included? The idea being that someone who IS capable of typing Cyrillic letters could come by later and enter the original text based on the images that would presumably be available.

It's true that I would also have this problem if I entered Japanese credits. Unless it is an online source where I can copy/paste the characters or if it is Kana only, I would have trouble entering any Kanji.

To be fair, in a case like that I'd say it might be better to stick with your own area of expertise. I've seen some credits sets transcribed where someone was obviously picking kanji out from charmap or something, and it's not pretty. Heck, I've even seen sets transcribed by native speakers -- with a list of errata in the original listing -- that *still* have typos I stumbled across while double-checking against screenshots. Not to to say I'm immune to this myself, having looked through some of my older sets...

[Ultyzarus]

I agree that we should stick to our areas of expertise. For instance, if I have a Japanese credits sets, I'd pass them to Jotaro (like I did on MG). Speaking of which, we could have a page somewhere on the site that lists people with specific sets of skills, so we could ask advice to those people for more complicated issues (not just post a question in a random place in the forums). There could even be a function to attribute credits and even a few points in a collaboration entry. In the case, for instance, that I provided Jotaro with a credits set, but he entered them in the database, Jotaro would get the points, but I would get something for providing the source. I'm not sure how it could be implemented though

[MZ per X]

If we simply left them misspelled in the credits listings and linked them all to the correct developer, we'd have a list right there where we could mark specific names as "misspelled" or "abbreviated", all on one page.

Yeah, no problem, I think. If a name is marked "misspelled" one time in the database, we can technically ensure that the user doesn't need to mark it again every time it shows up.

Perhaps there could be a flag for a credits set to denote that it's translated, but the original text is not included? The idea being that someone who IS capable of typing Cyrillic letters could come by later and enter the original text based on the images that would presumably be available.

Good idea, I added such a flag to the table for credit sets.

Speaking of which, we could have a page somewhere on the site that lists people with specific sets of skills, so we could ask advice to those people for more complicated issues (not just post a question in a random place in the forums).

It wouldn't be random, but a specific internationalization / translation board in the forums. That should be sufficient, shouldn't it?

There could even be a function to attribute credits and even a few points in a collaboration entry. In the case, for instance, that I provided Jotaro with a credits set, but he entered them in the database, Jotaro would get the points, but I would get something for providing the source. I'm not sure how it could be implemented though

I'm also not sure, seems somehow difficult to implement.

[jotaroraido]

Speaking of which, we could have a page somewhere on the site that lists people with specific sets of skills, so we could ask advice to those people for more complicated issues (not just post a question in a random place in the forums).

It wouldn't be random, but a specific internationalization / translation board in the forums. That should be sufficient, shouldn't it?

Or perhaps on the specific thread/talk page for the game? Though I agree that a translation request/help forum would certainly be handy.

There could even be a function to attribute credits and even a few points in a collaboration entry. In the case, for instance, that I provided Jotaro with a credits set, but he entered them in the database, Jotaro would get the points, but I would get something for providing the source. I'm not sure how it could be implemented though

I'm also not sure, seems somehow difficult to implement.

Perhaps it would be possible for, say, Fred to upload the attachment to the game and create a dummy credits set with no text. Then I come along and transcribe and translate it. Both of us get the appropriate credit that way.

I think we'd do best to avoid any type of points/achievement system that isn't sufficiently divorced from the database itself. The DB itself should be kept as pure as possible, with straight user attribution and change history as far as such things go there. That said, such contribution measurements make a good "gamified" incentive to contribute for a lot of people, but it should be something implemented on the client side based on the attributions, or possibly the change history.

[jotaroraido]

On a general note, we need a capable free text editor with database links (similar to the moby tags) not only for credits, but also for descriptions, reviews, walkthroughs, etc., so time spent on this is definitely not wasted. Maybe we can fork the technical basics from MediaWiki or somesuch.

Considering this a bit further, I've come to think that a good way to handle this would be something like the message entry system on these very forums. You can type in (or paste) raw text, type the tags and such yourself, or use designated buttons to create tags. Different types of fields would have a different subset of tags available -- it wouldn't make sense to link to a game from a credit listing, for example, while a review or a description wouldn't offer transliteration options.

With the raw text entered in the field, you could highlight, say, a role name, then click the "ROLE" button and it would give a small popup asking for the role type -- click on "programming" and it generates the needed tag syntax automatically.

Or for something a bit more involved, maybe you have a group of names under a role. Highlight the whole lot of them, then click the "PERSON NAME" button or whatever, and it goes through the names one-by-one (it could perhaps delimit individual names by line breaks, tabs, or multiple spaces), searching for existing developers that match. All matches would be listed in a pop-up along with links, along with an option to search for a different name (if, for example, the name is misspelled or a particular alias is not on file), or create an entirely new developer. The same could be done for companies, as well as for adding translations to text.

Obviously, there would also be the option to enter or edit the tagged text directly, allowing the option for offline tools to streamline or customize the process.

Anyway, I'm not in any way a web developer -- my knowledge of such things ends with HTML circa 1997 -- so I don't know how difficult it would be to create something like this. Thoughts?

[MZ per X]

Speaking of which, we could have a page somewhere on the site that lists people with specific sets of skills, so we could ask advice to those people for more complicated issues (not just post a question in a random place in the forums).

It wouldn't be random, but a specific internationalization / translation board in the forums. That should be sufficient, shouldn't it?

Or perhaps on the specific thread/talk page for the game? Though I agree that a translation request/help forum would certainly be handy.

Doing this on the game's page would be quite random, wouldn't it? If we had a central place for such requests where the capable people hang out, the chance of these issues being solved would be much higher, I think.

Perhaps it would be possible for, say, Fred to upload the attachment to the game and create a dummy credits set with no text. Then I come along and transcribe and translate it. Both of us get the appropriate credit that way.

Urgh, we had to extremely fine-grain our contribution and approval process for this. But that's more of a technical problem I think. Let' see what the developers come up with once they're implementing the contribution / points system.

On a general note, we need a capable free text editor with database links (similar to the moby tags) not only for credits, but also for descriptions, reviews, walkthroughs, etc., so time spent on this is definitely not wasted. Maybe we can fork the technical basics from MediaWiki or somesuch.

Considering this a bit further, I've come to think that a good way to handle this would be something like the message entry system on these very forums. You can type in (or paste) raw text, type the tags and such yourself, or use designated buttons to create tags. Different types of fields would have a different subset of tags available -- it wouldn't make sense to link to a game from a credit listing, for example, while a review or a description wouldn't offer transliteration options.

Agreed, that's what I imagine, too.

Obviously, there would also be the option to enter or edit the tagged text directly, allowing the option for offline tools to streamline or customize the process.

Yeah, that's important, as everybody can then have his / her own way to contribute such stuff.

Anyway, I'm not in any way a web developer -- my knowledge of such things ends with HTML circa 1997 -- so I don't know how difficult it would be to create something like this. Thoughts?

Dunno what's possible in HTML5, but JavaScript or such should be sufficient.

[Ultyzarus]

I've been trying my hand at some credits in MG (entered the few thousand names for Diablo III: Reaper of Souls) and I thought that it would be useful to not only keep the credits screenshots or scans used in backup, but to have a feature to tag where a specific name appears in it. I think this would be quite useful for determining if a developer should be merged / split, or just to see how it was spelled in a specific credits set.