RFC: D_D Translation tools

RFC written in collaboration with @Tao and @Alex1237

TLDR; Check out this video and let us know your thoughts!

What

A suite of translation tools that facilitate and incentivize the community to create high-quality, localized text resources in Web3.0.

Why

  • The DAO has a large, rapidly increasing number of wiki pages, forum posts, documents etc which are written in English. To be truly inclusive, we should translate all these resources into multiple languages accurately and on time.
  • The current translation process is largely manual and not scalable:
    • So far we only have translations for newsletters. The current process involves downloading the English doc, uploading to DeepL, copying the machine translated doc to Google docs then applying manual corrections, then finally post the doc link to newsletter
  • The current way relies on volunteer work which does not provide enough incentive for the general community to participate in and produce high-quality translations
  • This could be a tool provided by D_D for other DAOs as well.
  • If we could get the economics right, the tool could even generate income for the DAO and everyone who works on the translations.
  • If we build it smartly, it may become a public good, widely used by other DAOs or individuals.

How

Stack: Browser extension + Firestore + DeepL + Arweave

See below for a demo of a quick proof-of-concept.

:point_right: Demo video

Operational cost

Future projects / ideas

  • Discord bots: translate Discord messages real-time-ish
  • Translation of images

Areas where we specifically look for feedback

  • How to incentivize individual translators? Not sure how this would work, but it’d be nice to be able to record individual contributions and reward participants with tokens or something.

    • From Tao: My personal idea regarding the incentive for translation.
      • In the Arweave ecosystem, Atomic NFT makes it possible to sign and own pieces of the public decentralized data. The translated article which is archived in AR is also one type of atomic NFT. Koii is the attention leaderboard; the creator will earn attention rewards any time a user clicks on Atomic NFT. Based on that logic, the translators did all the work and cost to translate the article and save it in Arweave as Atomic NFT, the more users click/read the article, they will get a reward token (It might not be much in dollar value, but in a decentralized and crypto way) Yield Farming for NFTs | Koii
  • How do we ensure the accuracy of the translations? We could implement an upvote/downvote system but it sounds tricky if a single article is being worked on collaboratively by multiple translators.

    • From Alex1237:
      • A magic idea. We divide the firestore document into several pieces(Firesotre items), attached to a smart contract. Before starting a correction, a translator needs to stake a small amount of tokens for a week. Every time when starting and finishing a correction, he/she signs as a mark. In the next 7 days, anyone can report mistakes, like missing and messy, to local moderators. Once misbehaviours are confirmed, a penalty will be done. If everything is fine, the contract releases the stake and grants a reward to the translator.
      • Prerequisites: Moderators should be fair-minded and familiar with English translations. A webpage for staking and unstaking CODE should be deployed.
      • Issues: How to link firestore with a smart contract? How to prevent others from messing up your translation?
      • A graph showing my naive idea

10 Likes

Its great to see this initiative. I’m deeply inside the NEAR ecosystem right now and multi-lang has come up in a few conversations for grants and dapps that they aim to fund. Most of your tech went over my head (I’m Product/Marketing) though.

I love the idea of staking > translating > review > unstake > reward. That’s brilliant. If the incentives are correct (reward) then you ensure quality translations. Would you reward other participants? eg: QA

Cautions/Questions:

  • Might be possible to game the approvals - if you anon the QA, teams of people could gang together. So this approach might fail if the group of people are small
  • How do you vet participants? Maybe take a lean approach (ie give small task to start)?

Can we spin up a Discord or channel so I can contrib on the Marketing, product, ux etc?

1 Like

@daodir Thanks so much for the feedback! Yes QA would definitely be rewarded as well.

Given the fact that everyone is on their own schedule and the contributors would be all over the world, one thought I have is that there would be no set roles – anyone could chime in at any time to do some translation or QA or whatever improves the final result. The translations would continually improve over time. I’m not sure though how this would integrate with other aspects of the project such as staking or approvals.

So maybe we don’t even vet participants and let the community decide? I’d leave these questions to the capable hands of @Tao and @Alex1237. :slight_smile:

I’ll create a Discord thread/channel maybe a few days later when we have some more discussion points here.

–

Also question: what kind of requirements for multi-lang do you see in the NEAR ecosystem? Is it mostly static docs?

I don’t think a static doc is really a thing given the velocity in web3 so keeping translations in sync or flagging the “slave” when the master is updated needs to be a thing. Looking forward to hearing more.

The conv in NEAR haven’t been anything more than “we need some team to do translations” and we have grants ready to go for that

On a side note… doesn’t Amazon or Google have translation that gets us 80% of the way there?

Yes, the translation is from DeepL, based on TechCrunch, “Its translation tool is just as quick as the outsized competition, but more accurate and nuanced than any we’ve tried.”
Personally, i did quick test using both DeepL and Google from English to Chinese. DeepL is more nature for native speaker.

From my perspective the two key valuable feature of this tool sets are:

  1. Experience translators should have more background knowledge and add more information/culture background in the article to help reader to understand. Like movie scripts translation.
  2. We can incentive the translators or any participate using the Atomic NFT model and Koii Marketplace (Yield Farming for NFTs | Koii). Basically, the more user click the article you translated/saved, the more attention token translator will get. We can involve more with them to identify a better reward system specific for translation work.

Thanks for your attention and interest in our plan.

–Would you reward other participants? eg: QA

In Developer DAO, most questions are answered by team members, moderators, and DAO members. Team members and moderators should handle questions on duty. How to incentivize DAO members is our target.
The challenging issue of Q&A is hard to qualify. We cannot follow the same way of Translation Reward. Building a separate forum or webpage for Q&A is not convenient and unnecessary. The best place is Discord with bots in my mind. But I do not have a sustainable idea now. I am not sure how to prevent cheating Q&A records.

–How do you vet participants? Maybe take a lean approach (ie give small task to start)?
I don’t want to qualify translators, which may discourage members. From my experience, the high-quality translation should not be required and set as a standard. Reaching entry or medium level is enough. Just spread DAO knowledge in an understandable way. Those willing to mess up will receive a penalty and reporters get rewards for their justice.

When I meant a separate “thing” for QA, I was meaning an interface to provide translation, not notifications. For notifications, yes, maybe Discord is fine.

For actual translation, the user should have an app. Friction must be minimized. I think OP was considering a Chrome extension.

Sorry if I misunderstood you

I think what @daodir meant by “QA” is “Quality Assurance”, not “Questions and Answers”.

1 Like

ohhhh, I see. I thought you mean Question and Answer :smiley: haha

Since NEAR already has grants for translations, does it make sense for us to collaborate and do a pilot project or something to test how the reward system would work? Do you know a ballpark of how much the grant would be?

Oh… haha… yes @lintonye . Quality Assurance

Well NEAR doesn’t have specifically set out grants for translations per se… its just something that was a discussion during my call with one of their team members. I suggest you poke around the eco, and then just apply with your project/product. Check out the fellowships as well at near.university.

Thanks. I just had a quick look at the fellowships which look interesting. However, I think what we are looking for here is a bit different. We would just provide a tool rather than doing the actual translation. I think the goal is to verify the reward system. It might be easier with the grants that NEAR might have for translators – or does it make sense?

Also @daodir since you are already involved in the community, do you mind having an initial discussion with key members and/or make an introduction?

Yes, that makes sense. Does this project need funding for building/testing the app? Or Once built, will it be free to use/install? Then someone like me could apply for a grant to setup a translation project.

I will make some intros yes… give me a few days

As other D_D projects, this tool is for public good and will be free and open source. There will be some operational cost (see above RFC), although minimal, I expect it to be covered by D_D treasury.

With that said, funding for building/testing the app would definitely be helpful. I’m not sure though how external funding would work in the context of D_D projects. Like, would we get some grants from NEAR and somehow this remains a D_D project? Maybe through some kind of partnership? I’d be happy to explore this within D_D and NEAR if you could make an introduction.

Awesome! I created a thread on Discord and we could use it for logistics discussions.

Let’s keep technical discussions here in this post for organization purpose.

I’m going to need your help to get into Discord as I was unable to purchase an NFT (some tech issue a while back). I tried to fix it but never got a response from the team. So now, I’m not a Discord member. You will need to invite me. My Discord is MJ#0001

hmm… I don’t think it’s going to work even if I send you the invite link. Right now holding an NFT is required to access the Discord server. This will change and new membership will be supported once the governance token is launched early next year. Check out this post: Governance Token Proposal - Request for Feedback

Can you not send a msg to one of the admins? I’m sure they can invite me on