A tool that reports mistranslations with just one click

[rakunavi]

Hello all,

The PDF-XChange Editor user interface is multilingual.

• https://www.pdf-xchange.com/languages/pdf-xchange-editor

I only know about Japanese version, but frequently used features are accurately translated. If we were to divide translation quality into the following three broad levels, the majority would be Level A

1. Perfect translation
2. Translation that makes sense but needs improvement
3. A translation that does not convey the meaning at all

However, when it comes to detailed features, there are some translated texts that do not make sense, and sometimes the meaning can finally be understood by changing the UI language to English. In terms of the three levels above, this is Level C. This is probably because translators who have never actually used the function will keep save appearances by using machine translation in order to increase the rate of translation on the surface. If it is a translation that no one understands, it would be better to leave it in the original English text.

My English skill is not good enough to speak for others, but considering that I gets replies from support people and others, I think that at least the meaning is being conveyed. In terms of the three levels above, my English level will be Level B. If it is level C, I apologize.

Currently, the following methods are available for reporting mistranslations

• Send an e-mail to support.
• Post in the forum.
• Report a bug using the Report Bug tool on the Help tab.
• Use the PDF-XChange Localizer Language Tool provided to translators to check the identification ID of each element in each dialog and report it.

However, each and every mistranslation is merely a mistranslation. It is an extremely small issue compared to a bug in the application. As mentioned above, the meaning can be understood if the UI language is changed to English, and many users will eventually give up when they consider the time and effort required to report and the administrative burden on the other side after reporting.

I have also reported what I believe to be application bugs on this forum as soon as I find them, but I have reported mistranslations only once in the past. That is not because mistranslations do not exist, but because there are too many of them. Even if I would like to point out a mistranslation, just writing a report that conveys the mistranslation location on the UI is a challenge, but the language hurdle prevents me from getting the momentum to report it.

I think that a tool that automatically reports mistranslations by simply clicking on the mistranslated part of the user interface with the mouse would improve the quality of the translation. The easier the reporting procedure is, the more feedback you can expect from the user side.

Also, if feedback is automatically given to the translator, it will lower the hurdle for the reporter to report the issue. This is because the more people involved behind the scenes after an issue is reported, the more hesitant the user will be to report it. Ideally, feedback should be provided automatically to the translator by the user who discovers the mistranslation, without any other human intervention.

Thank you for taking the time to read this message.

Best regards,
rakunavi

[TrackerSupp-Daniel]

Hello, rakunavi

Thank you for the excellent suggestion, Revamping the translation process is something that we have been working on, and while it is not a top priority item, it is in progress. I have brought this idea to our Dev team, and while they like it, they have also identified a few key issues with this that prevent us from going forward with it at this time.

For now, any translation errors which are reported (which to be entirely honest, we only see 2 or 3 reports of the like every month, at most) can be sent to us via email or using the "report a bug" tool, so that we can directly fix the issue, or pass the suggestions along to the relevant volunteer translator for review.

Kind regards,

[rakunavi]

Hi Daniel, Thanks for your comment.

Even just opening a dedicated submission form would probably increase the number of reports. A simple form that allows users to submit a screen capture of the translation with errors marked with a circle or something similar would be effective. Optionally, it would be good to include suggestions for alternative translations.

Whether it is an application bug or a translation error, it should not take long to first become convinced that it is a bug or error. In the case of an application bug, you need to determine the cause, whether it is a mistake in your usage of the application or whether it is caused by the environment in which the application is used. In the case of a translation error, however, you only need to check the difference from the original English text, which should take less than a minute.

The current procedure for reporting translation errors can be divided into the following three steps

1. Recognize a translation error.
2. (Optional) Consider alternative translations.
3. Think of a message to convey the the translation error.

In my case, the second and third steps are the bottleneck compared to the first. It is easy to point out the mistake, but it is another matter to figure out how to correct it. Therefore, if we could complete the report immediately by just going through the first step, the number of reports should increase dramatically.

As I write this message right now, I can determine the approximate content of the message in a few seconds, but when it comes time to actually put it into words, it takes a considerable amount of time. Almost no one would spend more than a minute just to point out a translation error. I think the current very low number of reports speaks for itself.

It is also important to be dedicated, both to the tool I mentioned in my first post and to the submission form I mentioned at the beginning. I would not want to post translation errors on this forum on a daily basis, because other users would hate me for it.

Thank you for reading this long message as in the previous one.

Best regards,
rakunavi

[TrackerSupp-Daniel]

Hello, rakunavi

Yes, as mentioned we do recognize that this is a problem, and one which we are looking at improvements to handle, but it is not a priority at this time. The reason being, the only way to speed up reporting process, beyond a simply bug report or email, would require us to refactor the UI with context sensitive actions tied to each and every individual tool, button, submenu, dropdown, preset, error message, etc (everything that can be translated), and reconfigure it to show the report translation error in these cases.
As you can imagine, this is no small undertaking, and would prevent a great deal of general development from occurring while the project is ongoing.

Then following that, as our Translators are volunteers, we do not have a system in place for them to directly receive reports, at the moment everything must go through us, which we then pass along to the current latest translator for the language in question, and then wait, sometimes only a few hours, and sometimes months, for them to receive the message and get back to us with the update.
As they are volunteers, they work on their own time, and we will not (and really could not) expect them to rapidly or reactively fix these items, although if any of our translators who do quickly react to such reports are reading this, we appreciate you very much!

Kind regards,

[rakunavi]

Hi Daniel, Thanks for your comment.

I understand that the translation of the UI is supported by many volunteers, and I greatly appreciate it. However, for example, in the <Author> field of the Japanese localization files, "Jungle, Inc" appears many times. I assume that it probably refers to the sales distributor of PDF-XChange Editor in Japan, but can it be called a volunteer in this case? Also, it seems that translators are provided with a free license of the PDF-XChange Editor, but even in this sense, it is difficult to say that everything is supported by volunteers. Of course, I don't think the free license is worth the reward for the effort, so in that sense it could be called volunteer.

In any case, I think it is important to create an overall system that does not put pressure on translators to translate somehow, even using machine translation, when they have never used a certain feature and does not know the meaning of the description. If they have never used a particular feature, even if the machine translation result is completely wrong, it will be released to the world without being proofread.

As I mentioned earlier, if a translator does not understand the feature, I wish he/she leave the original English as it is. Although the accuracy of machine translation has improved considerably, the results are still quite poor for short sentences. Since the text used on the user interface is often short, the translation results tend to be nonsensical. I am very concerned that just a few irresponsible translations could lower the reputation of the entire application.

Thank you so much for your continued support.

Best regards,
rakunavi

[TrackerSupp-Daniel]

Hello, rakunavi

If we had the people with the skills to manually validate each unique language which is submit to ensure that cultural and language specific idioms where preserved in the translation and that the translation is clearly not machine translated in some capacity, We would be able to perform the translations in house. Unfortunately, we do not have language specialists in house.

Regarding your note about the Japanese translation specifically, Yes last I have heard our japanese translator is indeed someone who works at Jungle Japan, however translators can use any moniker they wish for the translation files they submit. They are not on our payroll, nor receiving any recompense beyond a complimentary license.
In any situation, the issue with relying on translators is only part of the overlying issue, the reason why this is not happening Right now is the technical requirements of a system to offer this in any form of a more user friendly way than we already do.

As mentioned before, it is something we are looking at, and it is something we wish to offer in the future, but we cannot, and will not be offering it in any short timeline. As there is no further discussion to be had on the topic, I will be closing this thread.

Kind regards,