Is there a way to do multi-language (3 or 4 instead of 2) translation memory matching? Thread poster: Adieu
|
|---|
Adieu 
Ukrainian to English
+ ...
I recently got a project from Language B into English that is a 90+% match across the board for something I already did from Language A into English. It is nearly word-for-word.
Well, it IS from the same project (from the same client), just slight localization differences in the source documents for different countries. And while cutting and pasting stuff from my own records and slightly adjusting it still makes the work go fast and easy, I wonder if there is some sort of "harmoniza... See more I recently got a project from Language B into English that is a 90+% match across the board for something I already did from Language A into English. It is nearly word-for-word.
Well, it IS from the same project (from the same client), just slight localization differences in the source documents for different countries. And while cutting and pasting stuff from my own records and slightly adjusting it still makes the work go fast and easy, I wonder if there is some sort of "harmonization" procedure, if you will, to get autosuggests for segments that closely resemble something you already did in another language.
[Edited at 2021-07-19 15:27 GMT] ▲ Collapse
| | | | Stepan Konev
Russian Federation
Local time: 17:00
English to Russian
| Are you serious about harmonization? | Jul 19, 2021 |
Say, you translated a text from Japanese into English. Now you translate the same text from Russian into English and you expect memoQ can recognize Russian in Japanese? It doesn’t work this way obviously.
| | | | Adieu
Ukrainian to English
+ ...
TOPIC STARTER
Not quite. I meant syntax and terminology.
Say, for example, you have a segment with no TM match in the current language pair, but the MT engine output is a 90+% match to a segment you confirmed in from another language into the target language. It'd be nice to have that close other-source-language translation shown as a TM suggestion with a separate color code.
It's theoretically feasible. And it makes a lot more sense than you'd first think, since a HUGE share of tran... See more Not quite. I meant syntax and terminology.
Say, for example, you have a segment with no TM match in the current language pair, but the MT engine output is a 90+% match to a segment you confirmed in from another language into the target language. It'd be nice to have that close other-source-language translation shown as a TM suggestion with a separate color code.
It's theoretically feasible. And it makes a lot more sense than you'd first think, since a HUGE share of translated information actually consists of re/back-translations of common sources and standard concepts. Most of the building blocks of legal contracts, textbooks, manuals, scientific articles, project descriptions etc. used around the world in all major languages have original languages that they were written in (usually English, sometimes Latin, Russian, Chinese, Japanese, Korean, German). EVERYTHING out there in certain topics is built out of building blocks written in a shared original source.
Obvious examples: IT, medicine, religion
Stepan Konev wrote:
Say, you translated a text from Japanese into English. Now you translate the same text from Russian into English and you expect memoQ can recognize Russian in Japanese? It doesn’t work this way obviously.
[Edited at 2021-07-19 16:06 GMT]
[Edited at 2021-07-19 16:16 GMT] ▲ Collapse
| | | | | Good question | Jul 19, 2021 |
Stepan, some languages are very similar. I sometimes get near-identical texts in all three Scandinavian languages. Plus, in Norway, there are two spelling conventions, and every now and again I get identical texts spelt the other way.
But it is a rather niche problem!
| | |
|
|
|
Adieu
Ukrainian to English
+ ...
TOPIC STARTER | I actually meant something slightly different | Jul 19, 2021 |
Specifically, an option for checking MT engine output against very close personal translation matches in the TM for the target language.... REGARDLESS OF THE SOURCE LANGUAGE.
And maybe also have it learn terminology likelihood outside of language pairs. Say, "OK Cancel Click" as more likely words than "Agreed Unsubscribe Press", or "Principal Investigator Study Trial" instead of "Chief Researcher Research Test"...not separately, but for ALL languages that you translate into English ... See more Specifically, an option for checking MT engine output against very close personal translation matches in the TM for the target language.... REGARDLESS OF THE SOURCE LANGUAGE.
And maybe also have it learn terminology likelihood outside of language pairs. Say, "OK Cancel Click" as more likely words than "Agreed Unsubscribe Press", or "Principal Investigator Study Trial" instead of "Chief Researcher Research Test"...not separately, but for ALL languages that you translate into English (or whatever language you translate into). ▲ Collapse
| | | | Stepan Konev
Russian Federation
Local time: 17:00
English to Russian
| Change language | Jul 19, 2021 |
Ice Scream wrote:
some languages are very similar. I sometimes get near-identical texts in all three Scandinavian languages I never tried that but in theory you can export your TM (Norwegian for example) into a tmx file, then change language code form "NO" to "SV" and import that tmx back to TM. This is how you can fake languages and use a Norwegian TM in a Swedish project.
| | | | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Is there a way to do multi-language (3 or 4 instead of 2) translation memory matching? | Wordfast Pro | Translation Memory Software for Any Platform
Exclusive discount for ProZ.com users!
Save over 13% when purchasing Wordfast Pro through ProZ.com. Wordfast is the world's #1 provider of platform-independent Translation Memory software. Consistently ranked the most user-friendly and highest value
Buy now! » |
| | Anycount & Translation Office 3000 | Translation Office 3000
Translation Office 3000 is an advanced accounting tool for freelance translators and small agencies. TO3000 easily and seamlessly integrates with the business life of professional freelance translators.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
