Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Show HN: Lost in translations – a fun app of recursive translations (amanjeev.com)
24 points by Amanjeev on Sept 30, 2015 | hide | past | favorite | 31 comments


"My hovercraft is full of eels" http://lost-in-translations.amanjeev.com/lost-translations/2...

[1] https://en.wikipedia.org/wiki/Dirty_Hungarian_Phrasebook & https://www.youtube.com/watch?v=G6D1YI-41ao


My first try: "I'm not going to buy this Album, you will be disqualified."

Then I saw this post, haha.


You get a special medal for using something from Monty Python.


Note: It seems like I am about to run out of the free quota (I see the usage as 754K characters already). I grossly underestimated.

Edit: If I run out of the quota, which it looks like I will, you can use your own API key locally by using this package: https://pypi.python.org/pypi/ulit/. A neat feature is that you can use Google translate API as well but it is not nearly as much fun.

Update: I ran out of the quota but have emailed Yandex if I can pay a few bucks to keep it running. It seems to have under 50 mil characters, you pay about $15.


Another update: I think the quota will be reset at midnight moscow time (it is Sept 30th today). So, hopefully in about three hours the quota should be reset and the app shall be usable again.


Same problem as http://translationparty.com/, running out of API quota. Sad since it's a funny idea.

Here's how Translation Party is supposed to work: https://tctechcrunch2011.files.wordpress.com/2009/08/forcebe... (similar to Lost in translations).


Sadly, Yandex has a limit for characters in their free API. A thing I noticed is that there is no free version of Google translate but that is not as much fun as well.


I tried some popular quotes, there is an example: "Whenever you find yourself on the side of the majority, it is time to pause and reflect." was transformed to: Every time you log on to the website of the Majority, it is Time to stop and think.


Seems like an entertaining, tech-enhanced, multi-lingual, solitary version of telephone [1].

FWIW: "The quick brown fox jumps over the lazy dog" becomes "In veloce Marrone volpe Pull lazy cane" [2]

[1] https://en.wikipedia.org/wiki/Chinese_whispers

[2] http://lost-in-translations.amanjeev.com/lost-translations/1...


>>Seems like an entertaining, tech-enhanced, multi-lingual, solitary version of telephone [1].

Everyone within the family I have mentioned this idea to has called it that.


The following goes through and comes out the end unchanged:

"We are all Human beings. The Earth is our home. It is our World. We have to learn to share, before it's too late."


If Amanjeev runs out of free quota, here's the link - http://lost-in-translations.amanjeev.com/lost-translations/1...


This thing of saving the translations paid off. So at least I wasn't as bad in predicting but I still failed to estimate the free quota's limitations.


Interestingly enough, the Italian translation is somewhat incorrect, as it says "before it _was_ too late".


I believe that's due to the comma in the original sentence that makes its meaning ambiguous:

We have to learn to share, before it's too late.

Could be either:

We have to learn to share before it's too late.

or

We have to learn to share, before, it's too late.

with the latter being 'it was' and the former being 'it is.'

Interesting that it manages to get all the way back to the original sentence in any case.


Using text: "Hello, i'm trying out this translation madness, I'm hoping these text won't lost."

Afterwards, it has corrected my grammar: Good Day, I am trying this translation Madness, I hope that these Texts will not lose.

http://lost-in-translations.amanjeev.com/lost-translations/7...


That's pretty cool :)

Maybe you could spin the waiting time by showing the intermediary translations as they go?


Thanks! Indeed. I am also working on allowing users to select the intermediate translations steps and the language of choice they want to begin with (per API).


That's funny, I wrote a hubot script that does something similar a little while ago:

https://github.com/okize/hubot-garble


Neat! We should join forces: Here is mine https://github.com/amanjeev/ulit


That looks great!

This doesn't really qualify as multiple discovery but it is funny how common it is for disparate people to have the same idea simultaneously.


I was always sure I am not the only one with this idea but somehow never could find a similar web project via Google searches and never looked on Github. So, my bad.


Reminds me of http://ackuna.com/badtranslator which was way better back when you could still use Bing and Google Translate.


Interesting. I tried googling this before I started making this app but could not find any. I guess my google-fu is not good enough.


I think this could use some non-Indo-European languages in the language chain for fun.

Also, looks like all past translations are available through URL modification:

http://lost-in-translations.amanjeev.com/lost-translations/1...

by changing 1000 to 999, 1001, etc.


Thanks. Yes I want to add the ability to have users decide the intermediate languages. And yes the past translations are available. Thats how i wanted it to be.


I managed to get "Sorry, language not supported by Yandex" error somehow. Is the app hitting to the API limits?


Yes. That is a the quota limit. Unfortunately, I never thought I'd run out since I never thought anyone would want to use it this much. 'Tis just a small experiment I did not invest much time in. Sorry.


Question: Would anyone be interested in the translations (as a plain-text dump) ?


Isn't this more about repeated iterations than recursive calls to itself?


I guess you are correct. I was trying to really find a decent phrase but I guess I failed.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: