QA tool for glossary checking, with character-based target text matching
Thread poster: Samuel Murray

Samuel Murray  Identity Verified
Netherlands
Local time: 10:31
Member (2006)
English to Afrikaans
+ ...
Jul 7, 2017

G'day everyone

Can anyone recommend a QA tool that I can use to check my translation against a glossary? All I need right now is to check a bilingual text (format doesn't matter) against a glossary (source term + target term) to tell me which target terms are missing.

However, there is a small catch: I want the target text matching to be character based. In other words, if the glossary contains "pressure = druk", and my source is "blood pressure measurement" and the target text is "bloeddrukmeting", then I want the QA tool to say "target term found", not "target term not found", and ideally highlight the target term so that I can see what the tool thinks it had found, so that I can quickly see if it's a false negative.

Does anyone know of such a tool?

My translation is a two-column table and my glossary is also a two-column table, both in plain text, but I can convert it.

Thanks
Samuel


Direct link Reply with quote
 

Samuel Murray  Identity Verified
Netherlands
Local time: 10:31
Member (2006)
English to Afrikaans
+ ...
TOPIC STARTER
Verifika Jul 7, 2017

Samuel Murray wrote:
Does anyone know of such a tool?


Verifika is capable of doing character-based target text matching, but it doesn't show me what the matches are, so I can't tell for certain if Verifika had matched an actual word or just a set of letters that happen to match the word.


Direct link Reply with quote
 

Oscar Martin
Spain
Local time: 10:31
English to Spanish
+ ...
Xbench Jul 7, 2017



[Editat el 2017-07-07 14:52 GMT]


Direct link Reply with quote
 

Riccardo Schiaffino  Identity Verified
United States
Local time: 02:31
Member (2003)
English to Italian
+ ...
That is possible in Xbench Jul 7, 2017

In Xbench you can do what you need by using a combination of power search and project checklists. Unfortunately, you cannot do that against a simple glossary: you need to create a checklist item for each single term.

The best way (AFAIK) to do that is to create the first entry (e.g. "pressure" in source and -"druk" in target with the power search activated), save that checklist, save the project in Xbench to a xbp file, then open the xbp file in a text editor, and replicate the entry for each of the terms you need to check.

...this is the summary version of the suggestion--I think I'll need to write a blog post (with appropriate screenshots) to illustrate how this can be done. I'll try to do that over the weekend.

EDIT
Actually... the above is not necessary in your case--Oscar is, of course, correct: you can just add your terms to a bilingual glossary in Xbench, and the program will find all segments that do not match the glossary entry, even when the target entry is within a word (so Xbench will find that "blood pressure measurement
" = "bloeddrukmeting" is correct, and will flag as incorrect any segments that have "pressure" in the source but not "druk" in the target)



[Edited at 2017-07-07 17:43 GMT]


Direct link Reply with quote
 

CafeTran Training
Netherlands
Local time: 10:31
CafeTran? Jul 7, 2017

Perhaps in CafeTran, using a stemming setting for QA glossary?

Direct link Reply with quote
 

Michael Joseph Wdowiak Beijer  Identity Verified
United Kingdom
Local time: 09:31
Member (2009)
Dutch to English
+ ...
No disrespect, but... Jul 7, 2017

Samuel Murray wrote:

G'day everyone

Can anyone recommend a QA tool that I can use to check my translation against a glossary? All I need right now is to check a bilingual text (format doesn't matter) against a glossary (source term + target term) to tell me which target terms are missing.

However, there is a small catch: I want the target text matching to be character based. In other words, if the glossary contains "pressure = druk", and my source is "blood pressure measurement" and the target text is "bloeddrukmeting", then I want the QA tool to say "target term found", not "target term not found", and ideally highlight the target term so that I can see what the tool thinks it had found, so that I can quickly see if it's a false negative.

Does anyone know of such a tool?

My translation is a two-column table and my glossary is also a two-column table, both in plain text, but I can convert it.

Thanks
Samuel


Apart from whether it's possible or how to do it, why would you want to?

Michael


Direct link Reply with quote
 

Samuel Murray  Identity Verified
Netherlands
Local time: 10:31
Member (2006)
English to Afrikaans
+ ...
TOPIC STARTER
Because I hate sifting through 1000s of false positives Jul 8, 2017

Michael Joseph Wdowiak Beijer wrote:
Apart from whether it's possible or how to do it, why would you want to?


Because I hate sifting through thousands of false positives. Because I use my glossary mostly for QA. In the glossary are words like "blood" and "urine", so that I can see if a translator that went before me had neglected to change "blood sample" (bloedmonster) to "urine sample" (urienmonster) in a fuzzy match.

I'm not suggesting that QA should replace manual checking, but it's like using a spell-checker: it's an extra set of "eyes".

And the reason why I would like to see what match the QA tool thinks it had found, is because false positives go both ways. For example, if the glossary target term is "jou" (the Afrikaans for "your") and the segment's target text contains "kilojoule" (the Afrikaans for "kilojoule"), then it would be perfectly clear to me (during a quick visual check) that that is a false negative, if "jou" within "kilojoule" is highlighted.


Direct link Reply with quote
 

CafeTran Training
Netherlands
Local time: 10:31
Did you have a look at CafeTran? Jul 10, 2017

Hi Samuel,

Did you have a look at CafeTran, like I suggested?

Screen Shot 2017-07-10 at 09.56.25

Perhaps Prefix matching suits your needs, perhaps you need to use regular expressions. Perhaps it's already enough to leave the box Whole words unchecked.

[Edited at 2017-07-10 07:58 GMT]


Direct link Reply with quote
 


To report site rules violations or get help, contact a site moderator:


You can also contact site staff by submitting a support request »

QA tool for glossary checking, with character-based target text matching

Advanced search






SDL MultiTerm 2017
Guarantee a unified, consistent and high-quality translation with terminology software by the industry leaders.

SDL MultiTerm 2017 allows translators to create one central location to store and manage multilingual terminology, and with SDL MultiTerm Extract 2017 you can automatically create term lists from your existing documentation to save time.

More info »
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 »



Forums
  • All of ProZ.com
  • Term search
  • Jobs
  • Forums
  • Multiple search