So, there's this known bug with DeepL plugin about start tag not having a matching tag etc. The same file can be successfully processed by MemoQ DeepL with only some errors (only several problematic segments are left unprocessed).
Is there a way to achieve the same in Trados? Whenever the bug appears, it leaves me with an empty file, which is a bit of a shame really, and clearly a shortcoming of the software.
If not, which easy-to-use-and-preferrably-integrated tag cleaners/cleaning features should I use, along with specific instructions, please?
Adrian Komiński said:"- the core Studio product needs to ignore the fact there is a DeepL bug and process the rest of the file anyway" - so is there some work in progress regarding this?
It's logged, that's all I can tell you. If DeepL didn't fail in the first place this would not be needed, the priority is not very high.
Adrian Komiński said:What about the cleaners?
I don't really know what you are referring to here?
Jesus, any signs of the fix coming? Another Studio update today and just ran into this lame bug again. 20 minutes of DeepL'n and an error at the end. This is not only a waste of time and nerves, it's also a waste of money for pointless PAID machine translation.
Freaking technology these days…
You're obviously blaming DeepL for this, but the stupid Studio could at least tell us exactly which segment(s) it has problem with, so that we can delete it from the document and move on.
Is this still a problem for you?
Time will tell if it is; your suggestion to switch to Iris from Solid has worked with that particular file. In the past I've also had problems with different .docxs, so if that was fixed too - I'm looking at the future with hope.
Adrian Komiński said: I'm looking at the future with hope.
So, my hope obviously proved naive. Just doing a huge project with 3 files, one of which was rejected by Studio for some reason (pretranslation error), and the DeepL plugin failed to translate ANY of its segments. Not only this! DeepL charged me for the full project, it seems. Also, pretranslating around 900 000 characters took 3 hours, which is a huge waste of time since I will probably have to migrate the project to MemoQ anyway.
I even requested a refund from them on a previous similar occasion, but they suggested that SDL is to blade for the faulty implementation of the plugin. It does seem true, given that I never had similar problems with MemoQ. Some (individual) segments would occasionally be omitted there, but without compromising the rest of the file.
Shame on you SDL, and get to work eventually. And no, I can't provide the file for testing, it's hush-hush confidential.
I mailed you yesterday as I had the same problem with the plugin. We looked at this today a little more.
Adrian Komiński said:I even requested a refund from them on a previous similar occasion, but they suggested that SDL is to blade for the faulty implementation of the plugin.
Interesting comment from DeepL. We tested the file you sent us with German (it fails going to Polish) in DeepL and it works perfectly well, so the problem seems to be related to the language pair.
Adrian Komiński said: It does seem true, given that I never had similar problems with MemoQ.
Hardly a sensible conclusion given the tools are completely different! Just because there is only one DeepL doesn't mean they are going to be blameless for any problem that affects one vendor and not another!
However, we are going to add some additional error handling and logging into the plugin and see if we can determine the cause so we can either fix it ourselves (if possible) or provide the information to DeepL to resolve.
Okie dokie, good luck with testing. The question is why didn't you add "additional error handling and logging" some half a year ago, despite reports (not only from me) of the plugin clearly failing? I'm happy you're doing something and puzzled as to why it's happening so late, given how the problem wastes our time and money.
Many users of different nationalities have reported the occurrence of the same problem in these forums, so it seems pretty obvious that Polish has nothing to do with the problem.
And yes, MemoQ is a completely different tool. I also forgot to mention the fact that MemoQ doesn't actually compromise the whole translation upon the occurrence of a single error, which Studio fails at to this date, and I have lost faith in it ever improving.
I am trying to help you here despite your aggressive and unhelpful atitude. I'm perfectly happy not to engage with you at all if you'd prefer? You know nothing about development or where problems may be occurring or how easy or difficult it is to fix.
Part of the reason you don't get all the help you're looking for is because I for one have no interest in engaging with you most of the time. You don't pay for support and I don't get paid for half the time I spend trying to help users with difficulties in using the software. I certainly don't need to be doing it and having to read the unhelpful commentary from you.
In fact we didn't even have to build a plugin for DeepL... we chose to do that because no 3rd party developers were interested, and neither were DeepL, but we wanted to be helpful. We also give it away, so give us a break and try to be more positive and stick to the point... it'll help us both.
Jesus Paul, I am literally dumbfounded… I don't even know if I should start by 1) explaining to you that customers can and should have expectations (such as expecting a problem to be solved within a year from reporting), and have the right to be pissed if this is not happening; 2) enlightening you about the revelation DeepL has been in 2017 to the translation industry and how you probably should be interested in ensuring compatibility with the service; 3) clarifying the type of your affiliation with SDL or 4) explaining that the basic single PC Trados license cost me the equivalent of a monthly salary of a senior supermarket cashier in my country, and I shouldn't have to politely ask you for providing basic support, not to even mention expecting quick and no-nonsense solutions to the number of present issues with the software, including malfunctioning preview, malfunctioning shortcuts, malfunctioning PDF conversion, malfunctioning DeepL, mega-slowdowns of the interface when switching to WYSIWYG.