-6

I am currently developing a tool to translate words on the fly with googles translate website. To better understand googles querys I have to analyse their GET parameters.

This is an example created with language detection on, result language de and text 'good' at www.translate.google.com

1. client   t
2. hl   de // Home language
3. ie   UTF-8 // Encoding
4. it   sel.124835
5. multires 1
6. oe   UTF-8
7. otf  1
8. otf  1
9. prev conf
10. psl en // 'Language from' before (when you change it)
11. ptl de // 'Language to' before (when you change it)
12. sc  1
13. sl  auto // Language to translate from, source language
14. ssel    3
15. ssel    3
16. text    good // Text to translate
17. tl  de // Language to translate to, target language
18. tsel    0
19. tsel    0
20. uptl    de

Question: what do the parameters I have not commented mean?

Kara
  • 6,115
  • 16
  • 50
  • 57
Peter
  • 149
  • 1
  • 7

2 Answers2

1

Do not scrape their webpage to hack together an API, they'll eventually shut you off, they're not big fans of that type of thing.

Here is a link to their API documentation for Google Translate.

https://developers.google.com/translate/

castis
  • 8,154
  • 4
  • 41
  • 63
  • I don't want to earn any money with it and it's mostly for private usage. Would still be glad if you guys could help me instead of warning me of it's dangers and recommending paid services. Thank you none the less. – Peter Mar 21 '13 at 22:43
  • The fact that you think warning you doesn't count as help tells everyone here that you're never gonna make it. – castis Mar 22 '13 at 00:31
1

If you're trying to reverse-engineer something like someone's GET parameters, then that's your goal, so do it! But a better option would be to use their supported and documented Translate API.

Jonathon Reinhart
  • 132,704
  • 33
  • 254
  • 328
  • You know this one is a paid service? – Peter Mar 21 '13 at 22:42
  • 1
    You know there's a reason for that, right? Go your route, help yourself, let's see how long before they ban your IP :) – Damien Pirsy Mar 21 '13 at 22:52
  • Oh when they do I don't really care. I use Curl to request the results. I don't suppose they have my ip. Still sad no one here tries to help :( – Peter Mar 21 '13 at 22:53
  • How can't you see this is a better help then what you're doing? People are showing you the right way, I would really call thah help – Damien Pirsy Mar 21 '13 at 22:58
  • Is this forum about helping people or about judging their intents from afar? I'm still glad you took your time to reply even if you did not answer my question. It still showed me I have to consider my questions more carefully and express my self more clearly :) – Peter Mar 21 '13 at 23:29
  • Um, yes they *absolutely* have your IP. It doesn't matter what client you use... that's how the Internet works. Have you ever looked at HTTP request logs? I'm certain they also have automated systems that keep track of the number of requests from an IP within a certain time range. And they *will* cut you off. Where that threshold is though, is unknown. If you figure all this out and keep your # of reqs/hr low, you might be OK. But I wouldn't advocate it. – Jonathon Reinhart Mar 21 '13 at 23:38