Our data is released under various Creative Commons licenses.More information
If you love this content, please consider a donation.
Menu
Maybe this issue has long been discussed (then please give me a hint where to look for an answer), but I have just discovered an annoying behaviour in Tatoeba.
If you enter an English "you"-sentence which then is (correctly) translated into three individual German sentences with "du", "ihr" and "Sie" respectively, and each German sentence gets in turn translated into a French sentence (all three different with "tu", "vous" and "Vous"),
then there should only be three linked pairs of German-French translations, i.e. "du"-"tu", "ihr"-"vous" and "Sie"-"Vous". However, when I look up one of the German sentences, all three French translations are listed and if I look up one of the French sentences, all three German translations are listed. This is apparently not correct, since the wrong translations are only linked over the English sentence node and should therefore not show up. Any remedy?
I just opened a ticket on the bug tracker for this: https://github.com/Tatoeba/tatoeba2/issues/1588
Thanks, here an example:
English: #3636227
German (correct) translations: #4132262 #4132263 #4132264
For each German sentence, not only the English translation is given but also the other two German sentences, which really sucks, because in German they mean different things!
Are you aware of the distinction between direct links and indirect links? A direct link, which is shown in black text, is created when a contributor creates a translation of an existing sentence, or an advanced contributor creates a connection between two existing sentences. An indirect link, which is shown in gray text, indicates that there exists a chain of direct translations between two sentences, but the two sentences themselves have not been marked as direct translations of each other. When I look at that example, everything seems to be fine. Could you take another look?
As a side note, I don't believe that French uses capitalization to distinguish between different uses of "vous" the way that German does to distinguish between different uses of "sie" (="they" or "she") and "Sie" (= formal "you").
Why are some translations in grey?
Grey translations are indirect translations. In other words, they are translations of the translations, and not translations of the main sentence (the main sentence is the sentence in big letters).
We display them because they can be useful, but you should be careful. Their meaning may differ a little from the main sentence.
https://en.wiki.tatoeba.org/art...tions-in-grey?
To build on danepo's reply, I have two questions for BRW:
(1) Were you aware of the existence of the Tatoeba wiki and of the FAQ page that it contains?
(2) If you had seen this description in the FAQ, would it have answered your question?
I ask because I'm not sure that the wiki is not sufficiently discoverable, nor am I sure that the wording is as clear as it could be. We could add an example, perhaps with screenshots.
(1) I was aware of it, but too stupid to consult it.
(2) If I'd seen on the page displaying sentences and translations a hint like "Beware! Only the black translations with blue arrows are direct, the grey translations with grey arrows are indirect and therefore to be taken with a grain of salt!" I would have never posted my message in the first place.
OK, I understand that you have agreed to display sentences and translations in a certain way. But I think it would really help to tell users right on the serach result page, for example, in which order the translations are presented or how reliable each translation is or you might even consider to let the users decide themselves, in which way the search results are to be displayed.
I am in no way trying to bully, I think Tatoeba is fantastic and I like it very much. All I am trying to do is suggest a few changes to make Tatoeba even more customer-friendly.
So all the best to you and the whole team!
You're not the first person who has been tripped up by this. Clearly, the best situation would be for us to have a user interface that makes the difference between direct and indirect links intuitively obvious, with no need for warnings, and no need for the user to consult the help. However, interface design is a more difficult task than it would appear, with many tradeoffs to consider. Feedback like yours is helpful in figuring out how to improve things.
The problem you had is something we tried to solve, but the solution is still an experimental option.
If you go to your settings (https://tatoeba.org/eng/user/settings) and select the option "Display sentences with the new design...", you still see that the search results look like this: https://gyazo.com/df924e72badea27efdf90b67b45c2031
If this is what you would have seen, would it have helped you more understanding that some translations are to be taken with a grain of salt?
That's more like it. Actually, it might be nice to have all the translations listed like this:
Direct translations (i.e. directly linked and therefore most trusted):
[...]
2. grade translations: (i.e. linked through one other sentence, less trusted)
[...]
3. grade translations (i.e. linked through two other sentences, even less trusted):
[...]
4. grade... (if any)
It might be interesting to see, what a sentence can become after a few successive translations into very different languages, especially ambiguous ones.
This has been thought and suggested before, the problem is that it's technically more difficult to implement.
For now, we can only provide the first level of indirect translations (that is translations of translations) with a reasonable response time. Displaying the rest would take too much time.
Ah, yes, that explains a lot to me. I thought that the displayed indirect translations included all levels, even the farthest removed. The term "translation of translation" in your experimental version could also be misinterpreted in that way.
Even with only two levels of translations displayed it would help a lot, if they could be arranged a bit differently.
Is it technically possible to display the indirect translations grouped under the respective direct translations, maybe indented like this:
>This is the main sentence YXZ.
>>This is the direct translation A.
>>>>>This is the direct translation 1 of A, indirect translation of XYZ.
>>>>>This is the direct translation 2 of A, indirect translation of XYZ.
>>This is direct translation B.
>>>>>This is the direct translation 1 of B, indirect translation of XYZ.
>>>>>This is the direct translation 2 of B, indirect translation of XYZ.
and so on...
This way, all connections between sentences would become immediately apparent.
And yet another suggestion (not directly related):
It would be rather nice, if someone could program a feature in Tatoeba to the effect that, if you hover with the cursor over a sentence-number, a popup shows you the sentence and if you hover over a sentence, a popup shows you its number (if not already displayed). This would certainly save a lot of time if you are looking for specific sentences.
Hello BRW and welcome,
It’s always good to get opinions from new contributors because you’re not too much "used" to the quirks of this website yet. :-)
I think you have a point here. Indentation could be used to better show the relationships between sentences. However, our structure is not a tree but a graph. In your example, let’s say there is a sentence C that is a translation of both A and B. It makes C a "translation of translation" of XYZ, but then where should we show it? Under A or B? Or maybe under both of them?
About your other suggestion, the contents of a sentence is already shown on hovering a sentence number (for example in your initial post in this thread). About showing the number when hovering text, can you explain in which kind of situation it would be useful to you?
Yes, I am new here and I know that you have been thinking about these issues ages ago, so my questions and suggestions must seem to you rather childish and arrogant.
I appreciate all the more that you answered my post.
Regarding my suggestion of indentation, I see your point.
It is exactly the graph structure, which intrigues me about Tatoeba. What a pity then, that the output of a search only yields a list of translations in two categories "direct" and "indirect", when the crucial information about how the sentences are linked is lost. As I see it, the only way to get this information is by following each individual link (or am I wrong?).
When I hover over the number of a linked sentence in the log-list (right had side next to the entry), there is no popup showing the sentence. This might be useful.
With "hovering over a sentence", I was referring to the list of translations themselves. If you hover over a # or a blue/grey arrow it says "show sentence details", here it could show the sentence number, so I could quickly collect a couple of numbers, instead of having to follow each link first.
Enough of criticism. I've been having a lot of fun with Tatoeba since I joined, and I'd like to thank you for all your help so far!