Hi Benjamin Wagner 

Tweaking the segmentation rules is not going to help here I'm afraid.  I think that, if this is the accepted way to write exponents in Korean, this is a number recognition problem, hence only the number itself has the blue line under it (segment #2):

If I add the plus symbol to the source (incorrectly I know) it works fine, although still not recognising the Korean of course:

I'll take this up with the development team and see what their view on this is.  Generally we use the underlying libraries of your operating system so if Microsoft doesn't recognise this as a number then we won't either.  So for now I can only offer you four workarounds.

#1

Search and replace the source text, either in the source file or in the SDLXLIFF (you can use the SDLXLIFF Toolkit for this), and add a plus symbol to the front of all your numbers that follow the convention you show here.  So make the source look like my example above.

#2

Search and replace the target text after you have completed the translation to correct the STR 20 for example, and make it STR +20.

#3

Use the regex autosuggest provider to place all your numbers as you type - http://appstore.sdl.com/app/regex-match-autosuggest-provider/412/

#4

Use the Terminjector to correctly autopropagate the numbers as you need them - http://appstore.sdl.com/app/terminjector/759/

I'll come back to you once I have some feedback from development.

Regards

Paul

Awesome! Thanks so much for looking into this.

I'll look into your suggestions in the mean time.

Same issue when working with numbers with commas. I'm surprised this hasn't been a bigger issue for people working with currencies.

Honestly, it wouldn't be so frustrating if Trados would just allow me to create a unique TU for this source text, but CTRL + SHIFT + Enter does nothing. I guess because the source text is so similar, Trados thinks it knows better than me. Ultimately, if something is in the translation of the TU I've created, Trados should respect that when it auto-propagates.

The biggest problem with using something like Term Injector as a solution is that I'm not the person preparing the final file that will be delivered. As multiple people may be editing these files simultaneously before they're handed off to be prepped for delivery, Trados needs to store the correct finalized text for each cell in the TM so that they can be consistently retrieved.

Hi! Your above suggestions only work in specific scenarios but are not a viable workaround for most instances in which this problem arises. It's been over 5 years, and my team and I are still regularly encountering this issue. Has there been any progress toward fixing it?

Kelly Zubair

Hi Kelly, if you want Paul to see this it would be good to include a link to his ID so he gets a notification email saying he's been named. If not, he won't see this as it's a very old thread among lots of other very old threads.

Better still, start a new thread describing exactly how the above happens to you and it will be noticed by someone at RWS.

To include a link to Paul's ID, type @ followed by pfilkin without a space and it will transform into a link Paul which will notify him by email that you've mentioned him.

(You can also copy and paste the link in the title of a message as I did above with your ID).

All the best,

Ali

Paul
Follow-Up Question:
My issue is exactly as described above by my fellow team member. Is there perhaps another way to pre-translate the file in Trados that does not require altering the source or target by hand? Many of these strings appear across several files that multiple people work on over significant stretches of time. So it is important that they pretranslate correctly each time that format appears rather than having to manipulate the source. Someone new to the project may not know this is an issue (and therefore would not know they need to make alterations), and it could introduce errors into our translations without our knowledge. I am also reluctant to revise these strings by hand in Trados or the finished file, since the file I am currently working on has literally hundreds of context matches with this issue. I have attached a sample image for reference. Any help addressing the root issue would be very much appreciated! Thank you!

reproduced

funny

생각할 수 있는 임시방편으로는
저 위에 "Paul"이 이미 기록한 것과 같이
"Find and Replace"이면 충분하다고 보이고

클릭을 여러 번 해야 하므로 번거롭다면
지극하게 간편한 임시방편으로는
이 문제에 전용으로 사용할 "Plug-In"을 하나 만들면 되는 데
("One Click" 즉, 클릭질 단 한 번으로 모든 수정이 완료됨)
플럭인 제작에는 비용이 필요합니다.
-플럭인의 사용성, 활용성도 제한적일 듯..

예를 들어..

이번과 같은품질관리 시나리오에서 "전용"으로 사용할 수 있는 플럭인을 한 개 만든다고 하면..

"Tip" 버튼을 누르면, QA 문제가 되는 세그먼트를 "필터링"한 후 그 세부 사항을 "하이라이트"로 표시합니다.
-조건은 소스에는 "숫자 증가"가 있으나, 타겟에 "플러스" 기호가 없는 경우 정도로 가정 해 봅니다.

이제

"Do it" 버튼을 누르면, 표시된 세그먼트에 대해서, "찾기 및 바꾸기"를 실행합니다.
-즉, 문제가 된 타겟에 누락된 "플러스" 기호를 삽입합니다. 

자동화의 단순한 한 사례일 뿐입니다. 
동일한 기능을 수작업으로 완벽하게 처리할 수 있습니다.
다만, 플럭인은 분량이 많거나 반복적일 경우에 처리하는 것일 뿐입니다.

"RWS" 측에서 좋은 해결책을 들고나오면 그것이 최상입니다.
그럴 일은 없어보입니다만...