Start a new topic

QA glitch or not?

The segment

Model of the year ; 

(there are two blanks, one of them a trailing space) has been translated to 

Modell des Jahres;

(without any blank or trailing space).


Now QA claims

  1. Inconsistent trailing space
  2. End punctuation

at the same time.


Shouldn't it be "Inconsistent trailing space" only? Because "End punctuation" is clear when "Inconsistent trailing space" is triggered, isn't it? Or is it better to have both warnings, just in case?



1 person has this problem

If CT was to check for the results of previous QA actions during the current QA action, when those actions are preformed all at once, the implementation would be much slower. Both warnings are true, and if you perform each action individually, they should call the respective warning. 

> Both warnings are true...
That's simply not correct.
This is not an end punctuation issue, it's a trailing space issue, period.

 

In my opinion, Torsten's example belongs to an end punctuation inconsistency as well. The program should report it during this check.    

I don't understand that point of view. In both cases the end punctuation is a semi-colon. How is that inconsistent?

 

So how would you call the following QA error?


Model of the year ;

Modell des Jahres;


Note that there aren't any trailing spaces in both sentences.

I wouldn't classify that as an error at all.
In this particular case, the Engflish sentence is simply incorrectly punctuated, the German correctly. It goes without saying that the translator should not reproduce this error in the target langauge.

This is also language dependent, e.g. in French a space betore a question mark or exclamation mark is obligatory, in English, no space is equally obligatory, so again, marking that as incorrect in the QA would not be helpful.

 

Well, I should have made it reversed:


Model of the year;

Modell des Jahres ;


It does help to catch such inconsistencies for lots of language pairs. I agree it may not work for all the punctuation cases in a specific language pair. Yet, this QA catch brings about more benefit than harm overall.

Yes, that seems fair enough. Except that as far as I can ascertain (I just did a quck test), this isn't currently caught by the inconsistent end punctuation QA. (Or of course by the trailing space QA.)

 

That's true. In practice, this language depended error (space + a punctuation mark) has been excluded just because of French. I was making a point about including the white space check along with end punctuation marks. It is not harmful to have it there.

My implicit question was whether this QA error cannot be simplifed somehow. In the end, better two false positives than one false negative. If Igor says that this would slow down QA too much, this answer is satisfactory for me.


this language depended error (space + a punctuation mark) has been excluded just because of French


A kind of optional QA "Sentence marks with errouneous spaces" (for ".", "!", ";", "?" etc., to find e.g. "word .", "what ?", "say , this", "(5) ,") would indeed be nice.


Login to post a comment