Thesis#

Juicebox needs to streamline efficiency and save ETH where possible in the bear market. There are opportunities to do this where translations are concerned. Felixander will put a system into place to efficiently translate important documents and website interfaces. Felixander is a good candidate for this effort as he has previous experience working with translators IRL, both in a managerial capacity and also as a translator.

Specification#

The translation workflow has recently come under some scrutiny because of what appears to be inefficient operational and bookkeeping strategies. An overhaul of the translation department to rely less on human oversight and more on procedural automated steps will both save ETH, save the room for human error, and ensure quality translations are happening expediently. Specifically, moving translations to Crowdin so that they can be automated and so that we can have more control over what specifically is allowed to be translated is essential. This would also include shutting down DeWork as a translation platform. Additionally, having a review-system in place to ensure that translations are checked for accuracy is needed. Chinese-language translation, which needs special care and expediency to service the large number of Chinese JB members, would receive priority attention. Lastly, a system whereby translators can receive work without them having to fight amongst themselves will go a long way in promoting a prosocial environment within the translators channel.

The workflow#

The main things Juicebox should be prioritizing for translation is the website front-end (mostly done) and the “Learn” section of the info.juicebox.money. Most of the front-end interface has already been translated, but most of the “Learn” section has not. These are the two entry points to Juicebox and should be translated in the major languages. We have a lot of the website translated, but should still consider adding major world languages like Hindi and Japanese (market research will dictate which languages to focus on). This is a one-time work project, not an ongoing workflow. My suggestion is that we hire one excellent translator from each of these languages to translate the entire front-end and “Learn” section, and then review that carefully with other translators of the same language. The “Learn” section is currently not translated in any language, and we should decide which languages we want that translated into.

Once that is complete, most ongoing translation at JB will cease until we are in a better market where justifying such ongoing payment makes more sense. The only ongoing work after the front-end is translated would be Chinese, so that we can further support the Mandarin-speaking members of our community.

Over the long-term, this will save a lot of the translation budget, which is valuable ETH in such harsh market conditions. Also, it will trend toward near-zero oversight, meaning that current payouts around translation can be reduced and eventually eliminated. In the short term, this system will cost far less than the current system.

Specific changes to be discussed:#

• Felixander will manage translations to ensure that only the most pertinent documents are translated and uploaded expediently
• Felixander will work to do an accounting of what current translation jobs are finished, incomplete, ready-to-be-published, etc., to make sure that we can maximize the work that has been done so far
• Felixander will give information to bookkeeping department regarding payouts and work completed, so that an efficient ETH disbursement workflow can take place for translators to be paid.
• Felixander will put a review process in place to ensure that translations are reviewed and that translations are high quality.
• We should decide which languages are the most important to be used to translate the JB front-end “Learn” section and main interface. The front-end has several languages already, but the “Learn” section should be translated at least into some of the major languages

Point of consideration for future translation jobs#

Some ongoing translation will happen from time to time, but the idea is that this work is not ongoing indefinitely, but rather project based. For instance, Felixander is working on some blog posts that are meant to market JB, or the current blog posts by JohnnyD that explain NFT tooling, should be translated into major non-English languages. This can be achieved easily through Crowdin on a per-need basis.

Requesting payout increase amount#

Truth be told, I’m not sure how much work this project will entail, and thus I am hesitant to assign an amount to it. I would like to get started on this, put a good system in place, and at that time try to determine the commitment required for the role. At that point I’ll write another proposal that is up for a vote.

The context#

This is a discussion proposal. The goal is for the community to discuss and come to agreement on the ideas in this proposal. If there are ways to take part of this proposal to improve existing translation procedures, then all the better. If we decide that this proposal should replace the entire existing translation department, then so be it. The goal is to have meaningful discussion openly and transparently to improve a system that does not appear to be working properly.

By the numbers#

The “learn” section currently has about has about 4,000 words in it between the Overview, Risks and Glossary sections (arguably the most important sections to have in other languages. That equates to a $400 translation job at $0.10 per word per language. This work would need to be review, which at the current rate of $0.05 per word would come out to $600 per language. If we are putting these into five languages, we would have a budget of about $3,000 to execute this task.

My blog articles (five article series) so far are about 4000 words in total, which means that translations plus review would come out to about $600 per language. If we are doing Hindi, Russian, Chinese, Japanese and French, that would come out to $3,000 in total to translate those articles into each of those 5 languages.

To put all this into perspective, the last cycle of translators’ budget was over $5,000, over $4,000 of which was for work that has zero utility for us and that is currently being audited. With that amount of money we could have nearly translated both the entire pertinent “learn” sections of Juicebox as well as most of the market-facing blog articles. Instead, we have not done either as far as I am aware.