For Cycle 63 #1640
Comments
|
ACK |
|
Thanks for the russian translations, I hope this brings new traders, and I am also glad to hear you'll do some dev tasks at bisq 2.
I think that in order to be in line with compensation request #1622 a compensation of 800 usd instead of 1000 usd that you're currently asking for. I can't compare if russian strings are better or required more refinement from machine translation, or maybe that other compensation request is actually undervalued, but I think it's important to let you know my thoughts about how much would I value your contribution. |
|
Hi @MwithM Thank you so much for your quick review and comments on the translations portion of my compensation request! I honestly was unaware that there was a wiki page and a formal process for doing translations, as well as that Transifex is an official tool available to use for the DAO which produces logs and reports, etc. Nevertheless, although my translations were done differently from the policy, the pull request was accepted and they were merged into the next release, which I think was the right thing to do because these were needed badly. Certainly, I will remember this information for any future translations and follow the process. That said, I agree that the DAO is entitled to a breakdown of the translations that were done in the same manner as Transifex report would provide, and that the compensation request is supported by number of words and a rate-per-word disclosure. Therefore, I have extracted the changes from the raw GIT diff logs, excluded all symbols, special characters, and any English words appearing in a portion of the Russian string (for example, "BTC", "Bisq", "Ctrl + ENTER"), and tallied it all up at the standard rate of $.07/word. Here is my resulting tabulation: Word Counts
Now, I want to emphasize that this effort was not about just running strings through an automated translation tool and getting the results into the properties file. If you look closely, the translations category are not even distributed. The objective here was always to make Bisq1 as welcoming and as usable as possible to a native Russian speaker with limited English ability, to promote consumer confidence and reputation in the product and brand. Therefore, a big part of the effort here was deciding on which strings were most important to be translated, as well testing and verifying in the application itself if they displayed correctly in the screen space available and were intelligible and understandable. The effort was therefore quite manual and time-consuming to produce the highest quality value for Bisq, with the translation tool (mostly Google Translate, but sometimes compared and contrasted with DeepL) serving only to provide a basic first draft. I manually reviewed each and every output, sometimes made hand-corrections, and usually re-translated back to English to make sure the resulting strings were reasonable and correct (and they weren't always). Sometimes abbreviations were used to make it work. Some existing translations too were re-done because I felt they weren't accurate. In one area (the Mediator's payout screen), I actually removed the Russian translations and put the English back in on the horizontal series of checkboxes, because the space was tightly squeezed to exactly fit the English, and such very limited space made the Russian words unable to be read (this was an exception to the general rule, as I figured if we have a Russian speaking mediator one day, they are going to have to be somebody who knows English pretty well too). I would also like to add that, from Bisq's perspective, value-added to the application should not be measured by how many words were translated (which is a simple time-effort assessment of the translator), but rather by how many strings were translated (the incrementing factor by which the application has actually been improved). That's something else to please keep in mind, in case a change of the policy is being discussed. That said, upon review of the Spanish translations compensation request #1622 which you referred to, and the data results above, I am revising the amount of my proposal allocated to the translations. I'm certainly happy to receive further comments, answer any questions, and provide specific examples if there is concern or anyone is interested. Thank you so much! |
|
I think that the transifex translation compensation rule does not really apply here because he did a much more profound and quality driven task, including checking the UI if layout breaks as well as fixing bugs with other translations (like bugs with single/double quotes). |
|
Transaction ID: 3f66218b66d851166de2903649c8f485201ead84d1b862a4feac1b592aa6e5fe |
|
Thanks for your explanation, I think it's good for future reference to know why 1000USD was asked for this task. |
|
Hello again @MwithM, Very much appreciate your thoughts and understanding. However, as @HenrikJannsen and myself mentioned, about half of the translations effort done here was coupled with development, testing, and some new design work. Therefore, I felt that what you were really saying (or meant to say) was that I needed to provide better separation of development work out from the translation deliverable in order to present this request. In particular, the apostrophe fixes that @HenrikJannsen and I mentioned, really is better classified as development work, especially as I even had to change a JAVA source file to fix a few instances correctly. So, I'm sticking with the $464 for translations on this CR, and I don't think there needs to be another CR next cycle to provide any further compensation of work done this cycle. It's okay, as there is some more work to do in French and Italian language files, as well as I have a plan to make the handling of apostrophes easier for the developers and translators going forward. On a policy note, I also would like to reiterate my belief that as compensation requests are supposed to be based on value/benefit added to the Bisq platform (as I understand it), not the value of the time and effort of the individual, utilizing "number of words" and a "rate per word" is not the right metric for judging compensation requests of translations, but rather it is the "number of strings" since translation of a string like, "Are you sure you want to burn {0} BSQ?" can be just as valuable to, if not more valuable, than translating a long cautionary string like, "You hard drive was detected to have less than 2GB remaining, and we really don't recommend you run Bisq like this, because 1, 2, 3... could happen ...". This is why I provided the change in the percentage of the Russian language file that is translated as one way to describe the value of my translation contributions to Bisq in this request. Thank you! |
|
Accepted. |
Summary
[1377.00]Contributions delivered
The Faster Payments System (SBP) is an inter-bank funds transfer service in Russia launched in 2019 that allows individuals to make personal payments using just a mobile phone number, usually for free or very low cost. The addition of this new payment method into Bisq is long overdue, particularly as over the past year or so, most people living in that country have lost access to the major centralized crypto exchanges from around the world (ex: Binance, Kraken, etc.) due to the voluntary decision of those organizations to withdraw from the market. The SBP bank transfer service has been a game changer over the past few years since its inception, and significantly rendered the expensive alternatives of card-to-card bank transfers and bank wires obsolete for making personal payments. Planning for the adding of this new payment method actually began in January, 2024 under Growth Issue #288, however no developer was ever assigned to implement it, and the feature became stalled indefinitely until I entered the scene to inquire on status, finished the discussion to resolve the design and outstanding issues, and finally stepped up to work on the code and perform end-to-end user testing. This undertaking thus involved the full development life cycle of adding a new feature into Bisq.
Previous to this effort, only 59% (1,585 out of 2,676) resource strings in Bisq were translated in the Russian language properties files. The rest were still displayed in English. Even the words "Buy" and "Sell" on the upper menu tabs were not translated! This meant that the typical native Russian speaker with limited English ability might have great difficulty or discomfort with using the Bisq application, which is rather large and rather complicated until you get used to it. Not understanding the features or functionality of the Bisq 1 application and how they work, someone very easily could be puzzled and ignore an important message in the pop-ups that may appear and lose money as a result. The application was not very welcoming or respectful to Russian speakers in this state. Now there are over 2,100 strings in Russian (about 79%), and most of the messages that users would see during normal use of the application are displayed in Russian.
The combination of these two enhancements to the Bisq 1 application, plus some miscellaneous bug fixes (for example, the "Same Bank" transfer method did not recognize "BANK OF AMERICA" and "Bank of America" listed on two different peers as matching the same bank) has delivered significant value to the platform. The changes to Bisq 1,which is still the bread and butter of Bisq, is expected to attract new users from Russia and Russian-only speakers living around the world as new and old people discover Bisq and tell others, increasing the user base and trade fee revenue for the DAO.
Contributions in progress
The Bisq 1 French and Italian language files have an "apostrophe crisis" that I encountered this cycle due to a technical requirement to "double-up" apostrophes on some resource strings but not others, which can be confusing to new developers, translators, and reviewers alike. Therefore, next cycle I plan to fix the strings in those language files or add new code functionality to the string resource loader
Res.javato end the apostrophe madness!Also, I will collaborate next cycle with @pazza83 and @skaunov, according to their availability, to refine and update their first draft of the new Bisq Wiki page for the new Bisq 1 payment method (SBP) that I delivered: Bisq Wiki: Faster Payments System SBP
Finally, @HenrikJannsen has invited me to participate in the long-term Bisq 2 development effort, to the extent of my ability and availability, so that's something else I may be getting involved with at Bisq.
The text was updated successfully, but these errors were encountered: