I'll write this in English so that Gengo can read it as well.
Lately a new error check was introduced in the translation workspace. You probably all have experienced this: Not only do you have to 'recheck mistakes' that are actually the English words in [[[brackets]]]. And more annoying for the Dutch: almost every correctly spelled compound (samenstelling) is mistakenly accused of being an error. That counts for the group jobs as well, which is even more drive-my-head-through-the-wall-annoying (that last word IS incorrect ;)), since you can't submit it anyway and have to rebuild the sentence and make it awful.
On that matter of group jobs: My last job the client forgot a lot of his [[[ in his original texts. So in my translation I fixed that error for him. No - can't do. Since that's 'an error'. So the software is forcing me to make mistakes because it thinks I made mistakes. Kinda controversial eh?
Anyway, let me know what you think!
I have encountered this problem one or two times, but I always have an option that reads something like "still proceed''? (It's in small letters in the popup box I think.) It isn't a major annoyance, but it does take one more action to complete the translation when it does occur.
Hi Jasper, I reported the compound issue to our engineering team and they updated our Dutch dictionary to the most recent version. In their post-update test, "samenstelling" was no longer being identified as an error. Next time you take a job, could you please check this and let us know if the issue is fixed?
The same seems to happen with special characters (e.g. ä, à, ü, ù, etc.), which are not recognized by the spell checker, so the words containing them get truncated. For instance "Gläser" is reported as "Gl" + "Aser", "liberté" is reported as "libert", etc. Of course, that's not an issue in English, but it is for pretty much every other European language out there...
At any rate, as Sander was saying, you can override the spell checker and the missing "[[[ ]]]" reports. The problem is that the reports themselves become practically useless (and just a waste of time to look at) if they contain lots of false errors, as in the case of comments in the source language left by the customer. I would suggest that the tech team exclude the text between triple brackets from the spell check and re examine the way that text is checked for consistency against the target, since I've recently experienced a case where, copy pasting the bracketed text from the source into the target, I still kept getting warnings about it. And of course, they should also definitely solve the issue with special characters, otherwise there's no point in having an integrated spell checker (I actually use my browsers spell checkers, or check the text in a word processor).
Lastly, why don't you guys just use Hunspell as a base? It's open source, after all, and should give your techies less headaches, I believe...
Thanks for the detailed response, Mirko.
As far as I can tell, our engineers implemented a fix yesterday so that words within triple brackets aren't spell checked. If you or Jasper notice this problem again in the future, could you please take a screenshot and send it to me at email@example.com along with the job number?
Re. special characters, thanks for reporting this. Could you please do the same as requested above next time this happens?
Hi Nathalie, I don't think that Jasper ment that the dutch word for compound (samenstelling) itself was the problem. Expressions are often 2 words in English and 1 word in Dutch. When this is the case, the Dutch word is called a compound. The spellingchecker will mark it red as a warning most of the time, which is very annoying. And yes, the problem still exists. As an example: "Holiday Time" will translate in 'Vakantietijd'. This is correct in Dutch, but it will be flagged a warning by the spellchecker as you will not find it in the Dutch dictionary. Compounds are very common in Dutch and the gengo spellchecker is useless for them. Jasper, can you confirm that this is the problem you were telling us about?
You're absolutely right Robert, both what you recap and what Natalia mentioned as fixed. I still experience the same problems, and they appear to be the worst in group jobs. And I have no idea what dictionary is being used, but e.g. the word 'app' (which shows up a LOT) is a good Dutch word (according e.g. to Van Dale), and I keep having to 'correct' it.
It's workable, just a little annoying ;)
Thanks for the extra info Robert—apologies for not understanding the issue correctly the first time and for my very slow response.
Do you have any idea why these words would be recognized as wrong in the Dutch dictionary?
Either way, as the issue is still unresolved, would you mind sending your explanation to firstname.lastname@example.org? They are best placed to communicate this to our engineering team right now.
Hello Natalia, I was wondering if this still is an issue at Gengo? Since compounds and English words are very often used in Dutch. Thank you for the answer. Regards, Jennifer.
@ jennifer.moesa: Natalia has since left Gengo and I'm the new Community Manager :)
We have fixed this issue, and haven't received any tickets about this for a long time. Also, from our testing it looks fine. Please do let us know if you are still experiencing issues though!
Pleased to meet you Megan. Good to hear this issue has been fixed. I will let the support team and/or you know if I encounter such an issue.