Saturday, June 26, 2021

Could it be illegal to adopt, edit or delete sentences in Tatoeba?

Original Wall post: https://tatoeba.org/en/wall/show_message/37243#message_37243


Recently, there has been a discussion around copyright in which there were claims that adopting, editing and even deleting sentences in Tatoeba could lead to some sort of copyright infringement. I would like to address these claims because, obviously, we should strive to not do anything that is against the law, and I think everyone deserves to have a good enough understanding of this topic to make their own decision on what they should do.

For the record, the discussion happened on an orphan sentence (#190202). The sentence was considered unnatural, was adopted, then edited into a more natural sentence. This is a routine corpus maintenance task that has been performed many times in the past without any complaints, but this time, it has been argued that this is not compliant with international copyright laws. There was also a suggestion to delete the sentence, but here as well, it was argued that deleting would be unlawful.

Why? Because all of this is violating the moral rights of the original author.

Let's go through each claim more in details.


(1) Adopting a sentence

It was argued that adopting a sentence is similar to claiming that you are the original author. After all, when you adopt a sentence, the sentence is said to "belong to" you and you are said to be its "owner". This would be a violation of the right of paternity (one of the moral rights). Because of this, it was suggested that we suspend the adoption feature.

Now it is important to note that what I just described is only an *interpretation* of the adoption feature. And as far as I know, the law does not care as much about interpretations as it cares about the *intent*.

If you adopt a sentence for the purpose of claiming that you are its original author when you are not, then there is an intent and that is clearly wrong. If you adopt a sentence because you want to protect it from being edited by others within Tatoeba, or because you want to fix a typo in it, then you are just taking advantage of the features that Tatoeba offers. There's no intent to steal away the credits from the original author.

Similarly, there was never any intent from Tatoeba to transfer moral rights to whoever becomes the new "owner" of a sentence. These rights are anyway not transferable so it would have been a vain attempt. The concept of ownership and adoption have been introduced at the beginning of what I would call the "version 2" of Tatoeba (cf. https://blog.tatoeba.org/2009/01/new-validation-system.html). In the context of Tatoeba, the word "owner" has been used to designate who is the user responsible for the sentence within Tatoeba. It was intended to refer to an internal mechanism. It was never intended to designate the owner in any sort of legal aspect.

Of course, just because we do something that can be misinterpreted doesn't make us immune to consequences. We can do bad things with the best intentions in the world, it's still going to be bad.

In our case, there was possibly a bad choice of word. And if we want to steer away from misinterpretations, we could think of using another word, like "maintainer" instead of "owner". Sentences would no longer "belong to" a certain user, but would be "maintained by" a certain user.

Ideally though, I think we should implement a feature to standardize the way we do attribution. Because at the moment, it is not very straightforward. Attribution is done with tags, with comments, sometimes even with the profile description. If we had a standard way to indicate the source and original author of a sentence, then it would become much less ambiguous that what we call "owner" is not the original author. In some cases of course the owner will also be the original author, but in some cases not.

I honestly don't know when that could be done, but in the meantime, I would say there's no need to suspend the adoption feature. It's a fairly important mechanism in our system and removing it is just not worth the trouble. Yes, it can cause confusion, but that just means we have to spend some time clarifying what we mean by "owner" and the problem should be solved.

(2) Editing a sentence

It was argued that editing a sentence and applying significant modifications to it violates the right of integrity (another moral right), and therefore, Tatoeba members should not be allowed to edit other people's sentences, except for minor corrections. For significant modifications, the contributor should be creating a new sentence and leave the current sentence untouched.

I feel there is here a misconception of what Tatoeba is or what Tatoeba does.

If Tatoeba had a contract with authors to act as a publisher of their sentences, then yes, I can see how that would be a violation of the right of integrity to let others freely edit the sentences. But Tatoeba is not the publisher of anyone's work. It is only the publisher of its own corpus.

The act of editing sentence is just a shortcut for managing the content of the corpus. Instead of doing something that would take two steps (delete + add), it is done in one step (just edit).

Taking this shortcut can have consequences on a functional level. Perhaps the new version of the sentence doesn't fit anymore with the translations that are linked to it. Perhaps the initial sentence was part of a certain list and the new sentence would not be relevant in that list. Perhaps the sentence had some tags which become obsolete with the modification of the sentence. So it is not always appropriate to take this shortcut. But it is allowed and there should be no legal consequences.

As said, "editing" is essentially "deleting + adding". There is no law makes it illegal to add a sentence in Tatoeba, as long as the sentence can be reused into a work that is published under CC BY. And there is no law that makes it illegal to delete a sentence from Tatoeba. This brings us to the last point...

(3) Deleting a sentence

It was argued that deleting a sentence is violating the right of withdrawal (yet another moral right), and therefore, every sentence should be kept unless the author explicitly asks for it to be deleted.

That is an incorrect understanding of the right of withdrawal. The right of withdrawal means that if you have published some work, you have the right to ask for your work to be unpublished and it should be unpublished upon your request. It does not mean that whoever published your work is forced to keep your work published until you said otherwise.

Just like editing, deleting a sentence also has consequences. It can create an inconvenience for anyone who used that sentence for something. Perhaps someone added the sentence to a list, and it's suddenly no longer there, making their list incomplete. Or perhaps someone is developing an application that used the sentence, but the next time they decide to update their version of the corpus, the sentence is not there anymore, and it causes a bug in their application. But this is more of a user experience issue. It is certainly not a legal issue.


Conclusion: adopting, editing and deleting sentences in Tatoeba does not violate moral rights. The features of Tatoeba are definitely not perfect and could be redesigned to better integrate with the laws, but they are not breaking the laws.

That being said, I'm not a lawyer, this is only my personal point of view.

What I can say however is that generally speaking, we try our best to provide an environment that is as legally safe as possible for as many people as possible to participate. As this case brings up international laws, it highlights that we should take into account the strictest laws that we know of, and provided that these laws are aligned with our ethics, we should use them as a baseline.

Of course, it is not possible for us to know all the laws of all the nations in the world, nor is it possible to comply every time. In some cases, we may deliberately choose to not comply, especially true if the risk is negligible for Tatoeba and becoming compliant would require to have a team of full time engineers and lawyers working on it for months. We unfortunately don't have such resources, quite the contrary. In such cases, I think it is understandable that we must leave the responsibility to each user to make sure to adapt their usage of Tatoeba accordingly.

This was a bit long but I hope I could help some people get a better understanding on this topic. And of course, if you disagree with something I said or if you have any information that can help improve the legal safety of Tatoeba, you are very much encouraged to share your thoughts and knowledge.

Tuesday, September 8, 2020

A summary of Kodoeba, our first coding event

This year we decided to organize a coding event which we named Kodoeba. The main goal: grow our community of developers.

It was a 10-week long event which was welcoming pretty much anyone who either wished to contribute to the internal code of Tatoeba, or to develop an external project that makes use of the Tatoeba corpus. We didn't have a very definite scope since this was our first coding event and we wanted to include as many people as possible.

We were 5 developers coordinating this event:

We had 27 participants at the start and 16 of them were deemed to have successfully completed the event.

We don't know yet when will be the next Kodoeba, but we definitely want to organize it again. We are happy we could help the participants learn something new and it was a good experience for us as well! We could see the various issues new developers run into when trying to get their hands on Tatoeba's code. We could also see how providing an API will become a more and more pressing matter as more and more developers integrate our sentences into their projects.

So until next time, we shall work on improving the onboarding of new developers and start slowly rallying developers who have been long waiting for an API. Then when the time comes, we will organize our second Kodoeba, where we will (perhaps) finally develop an official API for Tatoeba...? At least that's an idea :)

Thursday, August 27, 2020

Our journey towards a mobile-friendly UI

Our first attempt to provide a better experience for mobile users started in 2016. That was a pretty late attempt, considering that smartphones were already very widespread by then, but better late than never.

We were faced with basically two options:

  1. Developing a mobile app. 
  2. Making the website responsive.

Developing a mobile app would have probably been the better option, but it was not an option we could afford. It would have meant dividing our efforts. The time we spend on re-implementing Tatoeba as a mobile app is time that we don't spend on maintaining and improving the existing website. So we went for the second option: migrating the existing UI towards a responsive UI and improving its design along the way.

This attempt was put on hold by the end of 2016, as we shifted towards other priorities. From this attempt, Tatoeba had some look and feel update, but was still not responsive.

Last year, however, thanks to Mozilla Open Source Support (MOSS) who granted us an award of $15,000, we were able to resume the initiative.

We first updated what was left to do with the look and feel, to make as many pages as possible visually consistent.

Then next (the hardest part), we re-implemented all the features from our "old" sentence component into our "new" sentence component, with some improvement and an updated design. Even though all the features were re-implemented, we could not completely switch away from the old sentence component yet. Throughout the whole process, our registered members have had the option to use either the old or new design. A majority have started to use the new design but there are still some issues in the new design we need to address before we can comfortably transition.

Lastly, in the past couple of months, we worked on making the landing page and the search results responsive. These two pages represent the largest portion of the traffic on Tatoeba so we figured it would be most beneficial for our mobile users to have these pages mobile-friendly first. The responsive landing page was deployed in July and the responsive search results in mid-August.

So finally! It took four years but we now have a responsive UI 🎉 

...Or at least partially. 

We will, of course, continue our efforts to redesign the rest of the pages, and also improve the speed. This might take another four years, who knows, but we will get there, eventually! You can, in any case, check our progress on GitHub. And if this is a topic which you wish to see progress faster, please consider contributing as a developer :)

We would like to thank MOSS again for their financial support, as well as everyone who helped with the development, with the testing. or who simply took the time to give us valuable feedback.

Wednesday, August 19, 2020

What's New on Tatoeba? - Your biweekly recap #23

What's New on Tatoeba will be published biweekly until the end of August.

KODOEBA EVENT

  • Kodoeba ended last week, on Monday, August 10th.
  • There will be a blog post to summarize what was achieved during the event.

UPDATES

  • The menu dropdowns are no longer hidden behind the private message sidebar. Thanks to Ricardo14 for reporting the issue and Trang for fixing it.
  • Various regressions have been fixed. Thanks to CK, Ricardo14 and jpear1 for reporting them and to rumpelstilzchen for fixing them.
  • The search results page is now responsive and nicer to view on mobile. Thanks to Trang for implementing this and to CK, gillux and rumpelstilzchen for their helpful feedback. 
  • The tooltips that display dates are now more accurate as they indicate the timezone (UTC). Thanks to AlanF_US for reporting the issue and to polkadotazn (one of our Kodoeba participants) for working on it.

ON THE WALL

LANGUAGES

As usual, we thank everyone who helped to translate the website, but perhaps in the future, we integrate something in Tatoeba to better acknowledge this work!



If you'd like to help to the development of Tatoeba, report issues, or are just curious, have a look at the GitHub repository.

If you want to help us translate the website to your language, you can join us on Transifex: https://www.transifex.com/tatoe...ite/dashboard/ and check this article on the wiki https://en.wiki.tatoeba.org/art...e-translation.

If you're especially happy with one of the updates, don't hesitate to personally thank our developers :) They're working in the shadow but they'll be glad to hear your feedback.



Original post on the Wall

Tuesday, August 4, 2020

What's New on Tatoeba? - Your biweekly recap #22

What's New on Tatoeba will be published biweekly until the end of August.

KODOEBA EVENT

  • We are in the last week of Kodoeba and we have five new issues solved on internal code since last time.
    • It is now possible to remove a sentence from a list from the sentence page. Thanks to Ricardo14 for reporting the issue and nummer_001 for implementing it.
    • There is a new visibility option for lists, in case you want your list to be shown on the sentence page to every user (and not just to yourself). Thanks to soliloquist for raising the issue and nummer_001 for working on it.
    • Unicode characters are now showing up properly in list names. Thanks to soliloquist for reporting the issue and ftumsh for fixing it.
    • You can now save create URL for templates from the advanced search. Thanks to Ricardo14 for triggering the idea and jpear1 for implementing it.
    • The sort options have been redesigned. Thanks to Trang for the design suggestion, AlanF_US for the help with the wording and Roverandom789133 for the implementation.

UPDATES

  • Mobile users can now change the position of the cursor in text inputs. Thanks to rumpelstilzchen for reporting this bug and Trang for fixing it.
  • When searching sentences sorted randomly, the results are no longer re-randomized when going back to a previous page. Thanks to AlanF_US for reporting the issue and gillux for fixing it.

ON THE WALL

LANGUAGES

  • Tatoeba is supporting one new language: Karakhanid.
  • Slovak has been enabled on production as a UI language.
As usual, thanks to all the members who helped to translate the website!



If you'd like to help to the development of Tatoeba, report issues, or are just curious, have a look at the GitHub repository.

If you want to help us translate the website to your language, you can join us on Transifex: https://www.transifex.com/tatoeba/tatoeba_website/dashboard/ and check this article on the wiki https://en.wiki.tatoeba.org/articles/show/interface-translation.

If you're especially happy with one of the updates, don't hesitate to personally thank our developers :) They're working in the shadow but they'll be glad to hear your feedback.



Original post on the Wall

 

Monday, July 20, 2020

What's New on Tatoeba? - Your biweekly recap #21

What's New on Tatoeba will be published biweekly until the end of August.

KODOEBA EVENT

  • We have two new issues solved on internal code since last time.
  •  Information about original sentences is now available in the "Base of Sentences" section of the Downloads page. Thanks to ftumsh.
  •  All the export files are split by language. Thanks to lbdx for the scripts that splits our exports, and to gillux and rumpelstilzchen for helping to optimize the solution.
  •  A few external projects are now ready to receive feedback. You are kindly invited to check them out and share your thoughts!

UPDATES

  • The vocabulary count has been fixed. Thanks to AlanF_US for reporting the issue and rumpelstilzchen for fixing it.
  • Searching sentences with a question mark at the end no longer returns an error page. Thanks to CK for reporting this and gillux for fixing it.
  • rumpelstilzchen solved a bug where the search was returning sentences in the wrong language. Thanks to him for the fix and to brauchinet for reporting it.
  • The landing page is now responsive. Thanks to Trang for implementing this and to everyone for testing.
  • Logs on our servers are now a bit less cluttered. Thanks to rumpelstilzchen for taking care of this, and to gillux for suggesting it.

ON THE WALL

LANGUAGES

  • Tatoeba is supporting eight new languages: Chinook Jargon, Southern Subanen, Ancient Hebrew, Pali, Baluchi, South Levantine Arabic, Southern Haida and Northern Haida.
  • Serbian has been enabled on production as a UI language.
  • Several other UI languages have been enabled on dev: Bulgarian, Filipino, Turkmen, Croatian, Slovak and Balushi.
As usual, thanks to all the members who helped to translate the website!



If you'd like to help to the development of Tatoeba, report issues, or are just curious, have a look at the GitHub repository.

If you want to help us translate the website to your language, you can join us on Transifex: https://www.transifex.com/tatoeba/tatoeba_website/dashboard/ and check this article on the wiki https://en.wiki.tatoeba.org/articles/show/interface-translation.

If you're especially happy with one of the updates, don't hesitate to personally thank our developers :) They're working in the shadow but they'll be glad to hear your feedback.




Tuesday, June 30, 2020

What's New on Tatoeba? - Your biweekly recap #20

What's New on Tatoeba will be published biweekly until the end of August.


EVENT


It's been now one month since our Kodoeba event started.


  • As far as the internal code goes:
    • Our participants have solved five issues, and seven others are on their way. You can find the details on GitHub.
    • Alexs has asked for feedback about the tags: https://tatoeba.org/eng/wall/show_message/35555. Be sure to share your thoughts if you'd like to see the tags in Tatoeba become more useful!
  • As for the external projects:
    • The other projects are starting to take shape, it's still too early to showcase anything. We'll have to wait until mid or end of July.


UPDATES

  • The search has been improved for languages using Arabic scripts, Indonesian and Tagalog. Many thanks to Yorwba.
  • The number of messages in the private messages has been localized, thanks to Ricardo14.
  • There's now a reset icon in the inputs of the advanced search. Thanks to Roverandom789133 for adding this.
  • We no longer unnecessarily store IPs in our contributions logs. Thanks to jpear1 for cleaning this up.

ON THE WALL



LANGUAGES

  • Rircardo14 posted some updatea about the progress of the translation of our UI: https://tatoeba.org/eng/wall/show_message/35518.
  • A new UI language has been enabled on the dev website: Serbian.
  • As usual, thanks to all the members who helped to translate the website!


If you'd like to help to the development of Tatoeba, report issues, or are just curious, have a look at the GitHub repository.

If you want to help us translate the website to your language, you can join us on Transifex: https://www.transifex.com/tatoeba/tatoeba_website/dashboard/ and check this article on the wiki https://en.wiki.tatoeba.org/articles/show/interface-translation.

If you're especially happy with one of the updates, don't hesitate to personally thank our developers :) They're working in the shadow but they'll be glad to hear your feedback.



Original post on the Wall