Open Registry of Game Information 

  • Credits

  • Talk about specific features of our upcoming online game database.
Talk about specific features of our upcoming online game database.

Moderators: MZ per X, gene

 #37347  by jotaroraido
 29 Oct 2013, 04:02
Okay, so I haven't seen any discussion of this around here yet, but as anyone who ever saw my work at Moby should know, game credits are my passion. One problem I always had with credits on Moby was the way aliases were handled. They had to be manually entered on a per-game basis, were far too easy to lose when updating data, and got to be a real mess when you had, say, a translated alias with a known real name attached to a translated company name.

One place that seems to have a pretty good grasp of how to handle aliases (and developer profiles in general) is VGMdb. Take a look at this profile for an example -- you've got a dedicated field for original characters if someone has a non-English name, a list of aliases (which is populated automatically based on what credits have been linked to them), links directly to companies and organizations the individual has been associated with, and even sub-entries if someone has particular distinct aliases that they use.

The works list shows what titles they've been credited on, what they were credited for, and even what alias they were credited with -- which makes it really easy to do research on a particular person. The actual release pages list what was actually credited, alias or not, but still linking back to the person's main entry. There's also a dedicated discussion thread for each developer, making it a simple task to discuss, say, splits/merges, known aliases, or bio details before they go into the main entry.

Another thing they do that I was always amazed Moby never had is profile pages for unknown aliases. Quite often people used the same (or similar) aliases throughout their tenure at a company, and even if their real name is unknown, it's still incredibly useful to have them linked together for research purposes.

If we'd had these features, the work I did on Moby to clean up, say, old Sega credits would have been much easier.

Anyway, that's my spiel. Thoughts?
 #37389  by idrougge
 31 Oct 2013, 19:12
All I know is that this has been discussed in the German forum and that everyone came to the same conclusion as you.
 #37390  by jotaroraido
 31 Oct 2013, 20:21
Well, in that case, I shall sit patiently and await the day when I may flood the system with Japanese names.
 #37434  by MZ per X
 05 Nov 2013, 21:50
jotaroraido wrote:Okay, so I haven't seen any discussion of this around here yet, but as anyone who ever saw my work at Moby should know, game credits are my passion.
First of all, a belated welcome to you, just glad to have you discussing with us. :D I'd like to jump on this thread to ramble a bit about people data, which we already discussed to quite some detail in German. I also adjusted our data model here and there to incorporate your examples.

1) People Data

Up until now, people data will be covered in two layers, a basic layer for the most common data and easy entry, and an advanced layer to enable more complex data to be entered into the database.

The basic layer contains the main name, as in the name the person became famous with. This can be a real-life name, nick, or both, all of these can be saved with their native spelling, too. Furthermore, birth and death dates are covered, nationality, gender, and a description. This basic layer provides the easy way for people data to get into the database, thus enabling credits listings.

The advanced layer is a personal event table. As the name suggests, this table can be used to save data about personal events like birth, death, marriage, gender change, degree, etc., all of them with a date, decription, and the possibility to enter a change of name(s) like at marriage. This facility can be used to fine-grain people data, enabling us to provide the possibility for full developer biographies / timelines, if need be.

Furthermore, there's links to company data with which we can map staff relationships like CEO/CTO/COO etc., or shareholder relationships. Other extensive linking of people data will arise when modelling the data of gaming magazines.

2) People Credits

As regards to people credits, the main data structure is the credit set. A credit set is every list of names that can be found, be it in a manual, in the game itself, on publisher websites, or whereever. We aim for preserving not only the data itself, but also the formatting of the credit set, if that's possible. The sets will need to contain a source, title, and language, and can contain a description, and whether it is a set for uncredited people, like that artist of the box art, or those Japanese programmers hiding in the source code.

The link between the credit set and the people will be Connection 5 in the wiki. As main data, this link will contain the job title as given in the source, the area of expertise, and standard role of this person (Data List 8 ), and his/her experience level (Data List 9). Furthermore, we will be able to enter the name(s) used for the person within the credit set, and whether this used name is a misspelling or abbreviation. With these data, we can then populate the people entry with all the names he/she was credited, and where.

Finally, we have the Connections 26 and 27 set up, where we can link a credit set to a specific release (usual case), or to a release group, if these credits are not release-specific.

3) Corner Cases

So much for the basics, let's take a look at some corner cases already solved.

1) More than one credit set per release.

One list in the manual, one list in-game, and additional credits within a readme file? Or different credits per language within a multi-language release? No problem, just enter every list as a separate credit set, and link to the same release.

2) Uncredited people

As mentioned above. These poor people get their own credit set with the flag "Uncredited". This set can then be linked to a release (e.g. for the cover artist), or a release group (e.g. for uncredited programmers).

3) Compilation credits

If a compilation contains a humongous credits list for every game bundled, it can be linked to the compilation itself, or split up in different sets per game, if that's possible, and linked to the game releases.

4) Credits changed by patches

Here we would save an additional, complete credit set, and link it to the patch entry.

5) Add-on credits

If an add-on overwrites the credits of the main game, and main game and add-on credits cannot be separated, this would be a new credit set, too, which would get linked to both the main game and the add-on.

6) Website / Youtube credits

These cannot be linked to a special release, so will be linked to one or more release groups.

Okay, now is the time to throw your corner cases at us. :) Or someone could think about the data lists 8 and 9, and expand or improve them. For instance, the standard roles for people per area of expertise are still missing. We agreed that these roles shouldn't be as detailed as in Mobygames, but a little more general.
 #37438  by jotaroraido
 06 Nov 2013, 02:08
Thanks for the detailed write-up. :) I don't think I 100% understand all the connections in the data model yet, but it seems like a lot of eventualities have already been accounted for.

My main concern, as I wrote above, is with the actual credits sets and their display, particularly with foreign-language credits and credits with lots of aliases. Let's say, for example, I come across the following credit:

原作 広井王子/レッドカンパニー

Romanized, this would be: Gensaku Hiroi Ōji / Reddo Kanpanī

Or translataed: Original Work Prince Hiroi / Red Company

Hiroi Ōji is the pen name of Teruhisa Hiroi (who goes by several other aliases as well), who is representing developer Red Company. I'm trying to wrap my head around how we could best display this sort of data while keeping it complete but still readable. At Moby, I would have entered it as:

Original Work (原作) Teruhisa Hiroi (Hiroi Ōji / Red Company [広井王子/レッドカンパニー])

...which is obviously rather cumbersome. With the model we have now, I'm assuming we could remove his real name from display, and just have "Hiroi Ōji" link straight to his person data page, but it's still somewhat cumbersome. Would we be able to have "Red Company" link to said company's data sheet while still being attached to the person credit? Perhaps there could be a toggle between translated and original text, like VGMDB does for tracklists?
 #37441  by MZ per X
 06 Nov 2013, 15:44
jotaroraido wrote:...which is obviously rather cumbersome. With the model we have now, I'm assuming we could remove his real name from display, and just have "Hiroi Ōji" link straight to his person data page, but it's still somewhat cumbersome. Would we be able to have "Red Company" link to said company's data sheet while still being attached to the person credit?
I imagine credit listings as being some sort of formatted text with links in it. This way, we could preserve the original formatting, if possible and worthwhile, while linking to people and companies behind the text, or enter transliterations for Japanese credits. Of course, we also could provide unformatted lists with all the data entered, for easier research.
jotaroraido wrote:Perhaps there could be a toggle between translated and original text, like VGMDB does for tracklists?
If we've got the above done right, it shouldn't be too much of a problem to let the user decide whether he/she wants to see the native spelling, the transliterations, or both. Or even toggling between the three options.

gene might want to throw something at me for all my technically complicated ideas, but this I have to risk. ;)
 #37453  by jotaroraido
 08 Nov 2013, 06:40
Excellent. I like the idea of having relatively free formatting... so long as usability isn't compromised. ;)

Anyway, something tangentially-related that I didn't feel it worth making a new topic for:

Will it be possible to have hierarchical relationships between companies? Just as an example, the company I work for is US company owned by a European company, which is itself a subsidiary of a Chinese company. The US division I work for itself owns another studio, which itself owns a further satellite studio. So it should be possible to have parent/child relationships to represent, in descending order of ownership:

Perfect World Co., Ltd. -> Perfect World Europe B.V. -> Perfect World Entertainment, Inc. -> Cryptic Studios, Inc. -> Cryptic North

I imagine that such parent/child relations between companies is already in the plans somewhere (though I couldn't find it when I looked earlier), but a further complication is those credited, named teams that nevertheless wouldn't count as their own company. For example, Namco's Project Soul, Tecmo's Team Ninja, or Taito's Zuntata. Certainly worth crediting, but one big problem with the way this was done at Moby was that if we credited, say Team Ninja for development, it wouldn't show up when browsing the main Tecmo listings, even though Team Ninja is wholly part of Tecmo. When a team name is only used for one or two games, or when the specific division isn't always known, crediting them in isolation actually makes the information *harder* to use.

In short: Will it be possible to have company entries that imply their parent companies when credited?
 #37454  by Ultyzarus
 08 Nov 2013, 14:10
jotaroraido wrote:Excellent. I like the idea of having relatively free formatting... so long as usability isn't compromised. ;)

Anyway, something tangentially-related that I didn't feel it worth making a new topic for:

Will it be possible to have hierarchical relationships between companies? Just as an example, the company I work for is US company owned by a European company, which is itself a subsidiary of a Chinese company. The US division I work for itself owns another studio, which itself owns a further satellite studio. So it should be possible to have parent/child relationships to represent, in descending order of ownership:

Perfect World Co., Ltd. -> Perfect World Europe B.V. -> Perfect World Entertainment, Inc. -> Cryptic Studios, Inc. -> Cryptic North

I imagine that such parent/child relations between companies is already in the plans somewhere (though I couldn't find it when I looked earlier), but a further complication is those credited, named teams that nevertheless wouldn't count as their own company. For example, Namco's Project Soul, Tecmo's Team Ninja, or Taito's Zuntata. Certainly worth crediting, but one big problem with the way this was done at Moby was that if we credited, say Team Ninja for development, it wouldn't show up when browsing the main Tecmo listings, even though Team Ninja is wholly part of Tecmo. When a team name is only used for one or two games, or when the specific division isn't always known, crediting them in isolation actually makes the information *harder* to use.

In short: Will it be possible to have company entries that imply their parent companies when credited?
I second that! Companies were always blurry at MG, sometimes, the main Studio was credited, some other time, the specific studio (like Atlus for Dragon's Crown, that I submitted, was changed for its parent company)

If every developer is put in relation with the company/studio they work for at a specific time (using the timeline described above), I think that hierarchy will be created by itself and that the whole Credits and Company listing will almost sort itself out.
 #37457  by Rola
 09 Nov 2013, 12:21
Relations between companies are not set in stone.
If you want to link them, how would you track changes?

I'd also want to point out the possible cases of company revisionism that crept into MobyGames. Things like Atari making claims of creating Alone in the Dark simply on the grounds that they acquired Infogrames few years ago (read: long after the game was made).

I'd prefer such relations only being mentioned in company description page - and not adding another database relation.
 #37458  by Patrick
 09 Nov 2013, 18:48
Rola wrote:Things like Atari making claims of creating Alone in the Dark simply on the grounds that they acquired Infogrames few years ago (read: long after the game was made).
Isn't it the other way around?
 #37459  by hydr0x
 09 Nov 2013, 19:11
Yep, Infogrames aquired the Atari trademark and renamed to Atari, so if they, under their new name, claim to have done Alone in the Dark, that's absolutely correct.
 #37477  by jotaroraido
 12 Nov 2013, 02:10
So I finally got around to taking a closer look at the current data model, and one thing jumped out at me as perhaps a bit too limiting: First Name / Last Name / Nickname. I can see several issues with a strict implementation of this right off the bat:

1. It doesn't account for middle names, suffixes, or any other type of name that has more (or less!) than a given name + family name.
2. It doesn't account for names in languages that use the reverse order; IE family name then given name. Many east-Asian languages do this, including Japanese, Chinese, and Korean. I'm sure there are others.
3. Specific to Japanese (at least, I'm not aware of it being an issue for other languages) is readings for names. VGMdb handles this with a separate field next to the native name. While in general it can be inferred from the romanized name, you're still going to wind up with ambiguities if it isn't specifically given. Additionally, it's not uncommon for credits to list hiragana readings, as opposed to the original kanji, and many names have numerous valid romanizations (Sato, Satou, Satoh, Satō...)
4. This may just be a matter of simplicity for display in the chart, but there doesn't seem to be support for multiple nicknames. It's not uncommon to come across a developer who has used a dozen or more different aliases, including variations.

Of course, these could all be accounted for already, and just not included in the chart for simplicity's sake. But hey, I need to keep up my reputation as being completely anal about this stuff. ;)
 #37478  by jotaroraido
 12 Nov 2013, 02:15
jotaroraido wrote:2. It doesn't account for names in languages that use the reverse order; IE family name then given name. Many east-Asian languages do this, including Japanese, Chinese, and Korean. I'm sure there are others.
Oh, and as an addendum to this, would we be using the native order (family name + given name, "Iwata Satoru") or the standard English order (given name + family name, "Satoru Iwata") for names in such languages?
 #37479  by Ultyzarus
 12 Nov 2013, 02:23
jotaroraido wrote:
jotaroraido wrote:2. It doesn't account for names in languages that use the reverse order; IE family name then given name. Many east-Asian languages do this, including Japanese, Chinese, and Korean. I'm sure there are others.
Oh, and as an addendum to this, would we be using the native order (family name + given name, "Iwata Satoru") or the standard English order (given name + family name, "Satoru Iwata") for names in such languages?
I would implement an option to choose between "Standard" name order and "Reversed" name order when first entering the person in the database. The display should be marked as such ie:

-Iwata Satoru (Surname first)
-Satoru Iwata
 #37497  by idrougge
 17 Nov 2013, 05:09
jotaroraido wrote: 2. It doesn't account for names in languages that use the reverse order; IE family name then given name. Many east-Asian languages do this, including Japanese, Chinese, and Korean. I'm sure there are others.
Hungarian.