They're passable for simple words like menu items, provided you have the editing... but they completely fail for dialogue, as they are unable to take context into account.

They are fail-hat at Japanese-to-English for a number of reasons. One of them is that anime/manga/game culture revels in using bizarre constructs for made-up names and such; this makes it rather pointless for a lot of things: as an item do you need to translate 薬 as “medicine” or “potion”? Both are valid, and a machine won’t be able to tell without unusually specific context.

And that takes us to the bigger problem: context. The Japanese language leaves a lot of context outside of spoken statements, whereas English requires a higher level of detail in order to make valid sentences. To translate from Japanese to English, the translator must fill in the proper details if it is at all possible (most of the time, anyway).

Not only that, but the MT often gets it just plain wrong. For example, I go to Translation Party and start with a daring statement: “You will never find equilibrium this way.”

E->J comes out this way: この方法で均衡を見出すことはありません。 “With this method, there is no discovering a balance”. Okay, not natural sounding, but that sort of makes sense.

Now back again: “To find a balance in this way.” What? That sounds like quite the opposite of what was said, and it’s a sentence fragment to boot!

Basically, you can use MT to get the gist of things, but it isn’t accurate enough to just manipulate into making sense. Things it might give okay results at, like item lists, would still probably be much more accurate if you consult a dictionary; the MT would give you enough false positives for a typical game that you’d waste the time looking them up anyway.

This is a romhacking website. The game was hacked and effort was put into hacking it. I think it should be allowed. I am okay with not hosting patches with minimal hacking such as palette hacking, but genuine hacking effort was put into this game. There are plenty of games here that have inaccurate or crappy translation, should we start removing those? Should we start monitoring the content of all translation hacks to see if they meet correct grammar standards and such.  What if I make a patch that changes the story of a game completely, but did this on purpose. Would it not be hosted because it does not have the exact story or content of the original game? As long as the hack makes note of where the translation came from it should be allow. If not, we should base all accepted hacks on the translators work and not the rom hackers and maybe change the name of the site to ROMTanslating dot net

What about some of "the scene" history then?

Link

All I'm saying...

I know, they're still machine translated though.

I think the real problem is here that RHDN doesn't stay true to its goals here.
This is supposed to be a central hub for people to discuss, share info and help each other. Also, I think this site can currently be happy about every news post as is.

We certainly don't need Naked Headless Mario Fights the Dick Nazis too badly in our database, as the majority of this community would not prosper from that (and might actually be turned away) and it might lead to other trouble, legally.

However, now Staff action basically resulted in a news post that pretty much said "We don't like it so we don't host it, please go somewhere else." That's not very central-hub-esque.
If somebody is not lucky enough to find a translator willing to translate the script, but has some actual work done, I honestly feel we should support that. I remember back when I hadn't found a translator for my projects and opened a personal project thread after posting a help wanted ad. IIRC it was received very well and I received emails from that publicity. I also created a Project Page which I semi-randomly update to show some pictures (even featuring cavespeak!). I finally found a translator and that really helped getting the anxiousness of failure off my chest.
So I think if there is something playable to show that will help attract possible translators and stir up interest, we should host that patch along with some screenshots and have that fuzzy feeling inside of us .
In any case I think it is wrong to deny people that make an honest effort support because a proof-of-concept script is machine translated. We all know that langfandood's translation was welcomed and got a newspost covering it. And nobody argued back then, denied hosting or even referred to another site entirely!! And that was not because of April 1st, it was because it was D and byuu. And that's it.
Also, I know the debate has been here when Naked Headless Mario fought the Dick Nazis the first time on RHDN. I remember that we wanted to see effort put into the game -- much like we don't host just any gfx-only hack that changes Mario to Luigi -- on one hand, but still remain scrupulous about quality assessment of hacks in general on the other hand. Now think about the many title screen hacks we host from another community member and now think about this translation. I feel we should not judge it poor quality based on a machine-translated script, moreso since this is a call for assistance!

@trumisery: Submit a In-Progress entry for the translation along with screenshots and a description of what you can hack and what you're looking for. Create a community profile with that and inform people of the translation on there as well. Then create a personal project thread and do the same. If you raise enough awareness and the game is not complete crap, somebody might be interested after all

cYa,

Tauwasser

I agree with that statement.

I think the discriminating factor between acceptance and non-acceptance should probably be something more like "did the author put any effort into editing ?".

Whether the source of a translation is the product of Babelfish, "Kodansha's Kanji Learner's Dictionary", or Japanese class 301 isn't a true measure of comprehensibility.  I've seen several poor translations which were done by hand... so much so, that one might wonder whether it was machine-translated (and not edited).  I've also seen a few where the translator's/editor's grasp of English was... shall we say, weak.  A machine-translated script, properly edited for continuity and flow, would be light years ahead of this.

Of course, I haven't seen the actual translation for the game under discussion yet, so I don't know whether my arguments support or refute the submission.  But then again, neither have the parties who decided not to host the translation here.

I think that the site administrators should have every right to approve or reject a submission, but I think it's poor judgement to reject a submission before it has actually been submitted.