tag:blogger.com,1999:blog-11719805.post114274549923416158..comments2023-11-03T01:45:11.288-07:00Comments on Lost Garden: The Translation GameDaniel Cookhttp://www.blogger.com/profile/10437870541630835660noreply@blogger.comBlogger13125tag:blogger.com,1999:blog-11719805.post-90993550753277178462008-04-30T14:58:00.000-07:002008-04-30T14:58:00.000-07:00Ack!! This page is unreadable, thanks to the vide...Ack!! This page is unreadable, thanks to the video link in the comments. Can't someone wrap that in an <a> tag?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-11719805.post-91654526169177496792008-04-08T13:58:00.000-07:002008-04-08T13:58:00.000-07:00Very interesting idea! I'm wondering how much of t...Very interesting idea! I'm wondering how much of this is specific to localization though; couldn't you apply the same logic to, say, programming? (This is essentially what open-source projects do, but how many open-source MMOs do you know of that could viably compete with WoW?)<BR/><BR/>I'm also detecting a chicken-and-egg problem here: you state that you need thousands of players as your existing player base in order to make a translation system viable, BUT you need to make the translation system integrated with your core game. So, you need to design a system that can't even be used unless your game is a moderate success. Sounds risky, in a genre that's already pretty risk-heavy.<BR/><BR/>A lesser but still important concern is that localization isn't just about words and culture, it can also be visual and audial. The "magic wand" icon that goes with most Microsoft "wizards" makes no sense in cultures where a Wizard isn't associated with Magic. How do you translate an icon? More difficult would be stuff embedded more deeply in the game: if your environmental artists make a yellow-and-black-striped barrier in the game to indicate warning, or use the color red to indicate damage, what do you do when those colors have different meanings in certain cultures? What do you do if your game includes voice acting, and it's important to not just translate the words but also the tone, inflection and quality of voice? You can only take things so far with your players at the helm.Ian Schreiberhttps://www.blogger.com/profile/03146360375570794401noreply@blogger.comtag:blogger.com,1999:blog-11719805.post-31715407515098649532008-04-02T21:45:00.000-07:002008-04-02T21:45:00.000-07:00It's a really interesting thing you're getting at ...It's a really interesting thing you're getting at here - making a rewards system to make players do things you want them to :)<BR/><BR/>Reminds me a lot of how Newgrounds works - sure, it's mostly a portal catering to 13 and 14 year old humor, but their system really WORKS. <BR/><BR/>What they do there is make the ENTIRE community one big meta-game. When a new piece of content is submitted, it is usually crap as they are flooded with content every day. Also, it might be stolen or abusive or illegal. Before any submission is scored it is first put "under judgment" until a critical mass of users have rated it. If you give it a "0" and so does everyone else and it doesn't make it past judgment, then you get rewarded.<BR/><BR/>Alternatively, if the submission passes judgment and you gave it a high score, you get rewards as well. So you're rewarded for giving accurate reviews.<BR/><BR/>You can also flag something for being abusive or illegal.<BR/><BR/>If a moderator comes by and looks at the flagged submission and determines it is plagiarized, outright pornographic, or illegal, it is taken off, and your "whistle" for reporting illegal activities is upgraded. If you cry wolf, it is downgraded until it breaks, at which point the moderators ignore you and everyone can see that you are a troll.<BR/><BR/>What this really does, is serve to make content that is good IN THE EYES OF THE COMMUNITY rise to the top, and crap sink to the bottom.<BR/><BR/>There's no accounting for taste, but it is true that what's on the front page of Newgrounds is representative of what most people on that site would like. <BR/><BR/>As you do these activities - writing reviews, uploading content, etc, you keep gettting experience points which upgrade your profile, with which you can purchase little pointless vanity goodies.<BR/><BR/>It's a pretty impressive system, and what you are describing here is basically an attempt to leverage this kind of community power for translation. I'd seriously recommend looking at newgrounds if you haven't already. If you're like me, you might not care for their taste, but their system is particularly robust and forward-thinking.Lars Doucethttps://www.blogger.com/profile/12304952359311395336noreply@blogger.comtag:blogger.com,1999:blog-11719805.post-81800230146637470802008-04-02T19:02:00.000-07:002008-04-02T19:02:00.000-07:00PS - GBGames' Google Video URL breaks the page lay...PS - GBGames' Google Video URL breaks the page layout in Firefox...Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-11719805.post-56358573293401242562008-04-02T19:00:00.000-07:002008-04-02T19:00:00.000-07:00I think this sort of system could work well, as lo...I think this sort of system could work well, as long as the players are made aware that the game has been developed for language X; other languages are an extra bonus, not core content. They should probably have the option of viewing [default_lang] only, [default_lang] with local translations, or local language only (falling back to [default_lang] if unavailable).<BR/><BR/>One important consideration is that translators should only be able to translate generic all-purpose text, or specific text they have personally encountered. Otherwise they'll have access to dialogue that could give away quest solutions or other game secrets.<BR/><BR/>Perhaps when they are talking to an NPC, a little "translate this" icon pops up next to the text. This logs that the player wants to translate this conversation later. They can continue playing, and at a later stage they are presented with a conversation log (containing only the text they saw) for them to translate.<BR/><BR/>Likewise, reviewers could have a "mark for review" button. But it might be better to allow immediate feedback for bad translations; "thumbs-up" and "thumbs-down" buttons would allow reviewers to quickly rate a sentence without interruption while still contributing to the overall translation score.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-11719805.post-40334630561219063212008-04-01T11:38:00.000-07:002008-04-01T11:38:00.000-07:00It kind of reminds me of part of this video about ...It kind of reminds me of part of this video about how Captchas are being broken simply by asking random people visiting your adult site: <BR/><BR/>http://video.google.com/url?docid=-8246463980976635143&esrc=sr1&ev=v&len=3091&q=captcha+site%3Avideo.google.com&srcurl=http%3A%2F%2Fvideo.google.com%2Fvideoplay%3Fdocid%3D-8246463980976635143&vidurl=%2Fvideoplay%3Fdocid%3D-8246463980976635143%26q%3Dcaptcha%2Bsite%253Avideo.google.com%26total%3D5%26start%3D0%26num%3D10%26so%3D0%26type%3Dsearch%26plindex%3D0&usg=AL29H234_Hn1c1O6gu6JqxqS-8eBjDwwtA<BR/><BR/>The part about captcha comes in early, especially at the 5 minute mark, but the entire video is fascinating.<BR/><BR/>Based on this video (and not just the adult site/captcha workaround part) and what you're talking about, it sounds like you could have a completely separate game to help translate your MMO, instead of using your MMO's users as the translators. You will be able to address the first bullet in your Limitations section if you can make this game fun, and if you do that, then the third and fourth limitations may be addressed as well.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-11719805.post-88087639290336659752008-03-31T17:14:00.000-07:002008-03-31T17:14:00.000-07:00Great comments! Re: Andy A more wiki-like system ...Great comments! <BR/><BR/><B>Re: Andy</B> <BR/>A more wiki-like system where motivated users tag content for improvement or suggest alternative phrasing is quite desirable. I love the organic self organizing, iterative nature of wikis and I can easily see how it might fit. One issue that I'm trying to solve with the review system is moving people beyond only translating the items that they are personally interested in working on. Some wiki-based translation tends to be spotty and there aren't good systems for encouraging broad coverage of a body of text. <BR/><BR/><B>Re: Context and continuity</B><BR/>Context is important. Ideally, the system would let users view the strings in place so that they could see how their efforts look. This would help with some of the string length issues that were mentioned. Continuity is also important. Seeing links to related pieces of text also matter. <BR/><BR/>I will say that I'm more from the NetHack school of narrative than the Final Fantasy school. Since long narratives tend to be throwaway content in most MMOs, I'd say that the NetHack school is winning. Regardless, your game design needs to be amendable to piecemeal translation if you want to crowdsource your translation. If your game derives large portions of its 'fun' from players absorbing novel-like narrative, then this system isn't for you. Of course, I might recommend that you get into a radically different business than online games. It gets mighty expensive when you stray from the idea that players are the ultimate content. :-) <BR/><BR/><B>Gaming the system</B><BR/>Gaming the system is quite possible, but unlikely. What it requires is mass collusion between random people. You would need a random selection of 10 people who all speak a specific language to agree to rate a translation on a set value. As long as the griefer population is less than ~20%, the review system should be able to determine review quality. <BR/><BR/>There are also some pragmatic feedback systems in place that reduce the chance of this happening: They end up hurting themselves by rejecting good translations. They end up hurting themselves and helping another griefer by accepting poor translations. <BR/><BR/>Of course, if a griefer population does dominate a particular language, there are statistical alarms and complaint systems that can help alert admins of a problem. Ultimately, there has to be a responsible, trustworthy human at the other end of the line. The goal of the translation system is to extend the capabilities of that person, not remove them completely from the service. <BR/><BR/>Interestingly enough, wiki’s and ARG’s do a lot of this already, albeit in a slightly less organized fashion. I took a creepy tone at the end since I don’t think we should ever forget that these game will always be mass manipulation of the player population. However, most of the people who volunteer their time feel good about the results. They are contributing to the community and the social machine simply gives them with an easy path to go travel down. <BR/><BR/>take care<BR/>Danc.Daniel Cookhttps://www.blogger.com/profile/10437870541630835660noreply@blogger.comtag:blogger.com,1999:blog-11719805.post-8727095611035703202008-03-31T05:46:00.000-07:002008-03-31T05:46:00.000-07:00Interesting idea. This approach differs from Wiki...Interesting idea. This approach differs from Wikipedia and the like, in part, because it doesn't allow improvements: one player does the translation, and it's either accepted, or it's not.<BR/><BR/>I wonder if a more successful approach would be to allow players to modify the content "live," as they see something that needs changing. This would give you a simple way to find new translators, when they volunteer themselves, and with some work gives you a less review-based way to monitor the "success" of individual translators. <BR/><BR/>Having many experienced players able to modify (or at least flag) surprising content would go some way towards keeping consistency within the game, even if it's not quite what you first envisaged.<BR/><BR/>It's slightly less clear how this would integrate into the overall game experience, but I'm sure you'll have a solution for that. ;-)<BR/><BR/>And once you've got that system set up, why not use it for the original language? Not only do you get free proof-reading, but it can't help but improve the community's sense of involvement.Andyhttps://www.blogger.com/profile/08505382749162975954noreply@blogger.comtag:blogger.com,1999:blog-11719805.post-76707707239490852742008-03-31T02:56:00.000-07:002008-03-31T02:56:00.000-07:00Hi Danc,In the "Limitations" part, maybe you shoul...Hi Danc,<BR/><BR/>In the "Limitations" part, maybe you should add "lack of context" : translating strings out of context can give very weird results and it's almost always preferable to give a translator the context in which the translation is being done.<BR/><BR/>I find your conclusion quite creepy but, unfortunately, it also has a good probability of ending up true...Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-11719805.post-88390624411811408062008-03-31T01:52:00.000-07:002008-03-31T01:52:00.000-07:00Yikes.I totally disagree with this article. The i...Yikes.<BR/>I totally disagree with this article. The idea is merely a "fan"-tasy, a purely theoretical view. I do love abstract, maze-like literature and the likes of JL Borgès, but we are talking about real life, here.<BR/><BR/>Sadly, what you are looking for is a huge army of free monkeys typing just one sentence on their tiny typewriters. This 1st army would be backed-up by a reviewing monkey army. In the end, higher ranked monkeys would reward both armies. Talking about a pyramidal society. Babel was a tower, not the Pharaoh's mansion. This would make a nice pitch for a looney Lucas point'n'click: "Let my people POGO the Great Monkey: The curse of the Babel fishbone" but it looks like you are looking for a real life solution. So this sounds creepy.<BR/><BR/>IMO, localization business is no hobby; MMOs are zillion-words long, ever-expanding games. This is definitely a full-time (even more!) job. <BR/><BR/>Good luck with your army and who knows? Maybe one day, I'll be enslaved in your simian typing pool.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-11719805.post-25754902017581332992008-03-31T01:40:00.000-07:002008-03-31T01:40:00.000-07:00Hi Danc,I am concerned that the feedback system yo...Hi Danc,<BR/><BR/>I am concerned that the feedback system you propose is gameable: if reviewer assign a score of 7 to all translations, they collect the reward and kudos, reward the original translator and boost their reputation without concern for the quality of the translation. The system relies strongly on the ethical stance of individual players and in-game guilds.<BR/><BR/>An alternate feedback mechanism is to have reviewers translate the player-contributed translation back into the original language, and then have players with no interest in translating to rate how similar the two texts are. This is vulnerable to negative gaming, where these non-translators score every translation as 'poor' in order to deny kudos to translators.<BR/><BR/>Perhaps removing the association of high scores with rewards would be viable: rewarding good translations is desirable, but requires an existing good translation to compare against…<BR/><BR/>If your player base is ethically sound and unlikely to engage in metagaming, then this approach would be suitable.falsedanhttps://www.blogger.com/profile/01052213510379373280noreply@blogger.comtag:blogger.com,1999:blog-11719805.post-7520543889747806512008-03-30T21:26:00.000-07:002008-03-30T21:26:00.000-07:00Intriguing, indeed. The idea of translation as a g...Intriguing, indeed. The idea of translation as a game in itself is certainly an interesting one, but I can't help but feel it would wholly unsuitable in the majority of cases.<BR/><BR/>The above article does address many of the potential issues with such a strategy, but here are two more:<BR/><BR/>1. Volume of text. Apparently, the amount of dialogue in Final Fantasy XI contained more text than the Bible itself. And bearing in mind that in-game dialogue is only a portion of the assets which need translation, you have to wonder if you'd ever reach critical mass in terms of translators able and willing to translate such a huge amount of text.<BR/><BR/>2. Uniformity of translation. Too many cooks spoil the broth, they say. Too many translators spoil the translation? This is also a problem that rears its head in the more traditional videogame localisation process. With such a huge volume of text, and unique worlds that often span dozens of games, how can you keep the text singing from the one hymn sheet? Not only in terms of tone and style, but also in more basic matters - things like character, item and place names? Usually, translation-memory programmes and exhaustive glosseries are used, but these would perhaps be unweildy to employ in-game.<BR/><BR/>3. Other concerns. Often, it's more than just language that needs to be localised. Whether it's an issue of censorship (Germany necessitates red blood is replaced by green), political differences (Football Manager was banned in China for recognising Taiwan as a country?) or cultural differences (characters' appearances are often hugely restyled for the Japanese market), there's a huge amount that needs to be localised - and this often involves artists, programmes delving into the game's code and assets to make these changes. So, no matter what way you slice it, it seems to me that localisation-as-a-game could only ever be part of a solution, and couldn't stand by itself.<BR/><BR/>That said, a very interesting article and one that offers plenty of food for thought.<BR/><BR/>PS. For a general overview on the translation and localisation process in the videogame industry, have a look at: http://brendan.tinnelly.googlepages.com/dissertation.docAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-11719805.post-24141603259983368902008-03-30T17:09:00.000-07:002008-03-30T17:09:00.000-07:00An intriguing concept. Especially the customer bu...An intriguing concept. Especially the customer built plot ideas.Unknownhttps://www.blogger.com/profile/13857778855289469682noreply@blogger.com