TED translating using volunteering translators, reviewers {I me mine} and approvers is a complex hogwash ecosystem that caters for the bad translators [driven by fear of bad translation] not the good ones. The ones who try the job and leave are the Chancers {here here}. They are by definition impatient and some of them might have stayed had there been a feedback. Chancers are 2 types: Morons and No-bs quickies. The Morons are not objectively inept but defined so by their reviewers. If your translation's review is more red than black, [the reviewer changed more than half of the subtitles] it means they think/see you should go back to school. The No-bs quickies do not generate red reviews but would not wait [it can be forever but is usually in months] for a review or approval and leave. As we can see a Chancer's fate depends largely on the reviewer. TED says the reviewers should be somewhat more experienced. Besides topical interest they have other 3 motivations by a) the teacher's pleasure of herding the pupils, making it red, b) seeing reviews as better language practice than translating itself and c) lubricating the publishing funnel. The reviewers are choosing according to the talk topic mainly. The lubrication motivation is nudged by seeing the community's stats at amara-tools but these are not emailed like new TED videos ready for subtitling, one must consciously seek them. The moronizing teacher motivation can transform into mentor-ships when both parts fit which is rare because the feedback comes late. A reviewer imagines the translator would recall the errors but after a week it is hard without links and examples so the most feedbacks do not do much. Mentorships may be nudged but willingness information on both sides is not tagged, collected nor publicly available with the exception of LCs who are so monopolising the mentor niche.
There are 2 pure types of retained [non-chancer] volunteers> the Bored bums and Hermionas. They have time to help since they simply have enough time [Bums] or can organise it [Hermionas] well. Bored bums are more flexible with rather varied productivity since Hermionas have their plans/diary. Their productivity stability is on a scale from Suns to Novas. Novas are fecund at a life's crossroad, looking around like a waitlisted student. We all are on a crossroads somehow, a Nova is there long enough to contribute like crazy and then the life moves on and they keep their 2 translations/approvals [LCs] per year. What crisis? All goes best it [in an unpredictable internet] can... funds are well spent, the community being enhanced and values promoted. Facts: TED says its raison d'etre is spreading of ideas. TED has 2 houses: 1st) TED conferences with exclusive congratulatory community & 2nd) the idea spreading machine: www.ted.com. TED thinks its translators serve the elite house but we, by our nature. should and do serve the spreading machine. The only tongues that have enough of the reviewers and language coordinators [LCs] are Spanish and Portuguese. Ted.com is a communication machine and TED translators' rules/culture is by one rule suits all indirectly hampering this communication. The rule is that the LCs have a privilege to control others´ work publication. Besides it is subjective it is also silly as good translators, like me, are not retained. Well, if TED translators were in the same business as ted.com [spreading of the ideas aka entertainment & education] it would have been different. TED would honour the users [global Internet public] with ease of approach and more of the translations. One premise says> a talk jumps in popularity after each new subtitles are added/published. Communication approach says: 'the errors happen, perfection is an enemy, the goal is to convey gist and create emotions'. The speakers, if they do not memorise, make mistakes. Some sentences are nonsensical or there is bad intonation or structure. It is OK, it is a talk, not a book. A translator sees the gist and conveys it... often with their own typos and mistakes... and here lies the problem of TED videos' translations... the way it copes with these mistakes... it gives them too much power and relevance [enemy] and is on a quest of eliminating them; whereas they can be used as communication enhancements and nudges. If one, as a new no-English user, visits www.ted.com he/she either googles on or leaves with nothing to watch. The language mutations are missing as well as searchable lists of subtitles. The good job/help TED can do is to help the non English users to their selected talks. UCD webdesign. Perhaps via a new portal link with set of alternative ways on how to get to talks of varied interests and tongues. Amara-tools shows a way here. TED translators to LCs funnel is like a real socialism organisation. The applicants write essays on how they will help so that the politbyro can approve of their ability to belong into the language community coordinating class with post editing privileges. Coordinating, subordinating? They should help the novices. The entry barriers can be lowered. Who are the novice volunteer translators? Mostly the learners of English [personal fitness], then lovers of some ideas spreading [inclusive fitness] then, secondarily, members looking for meaning in a good-doing club. Coordinators should probably groom the novices so that they feel more of the belonging motivation. I would prefer/like to see a motivation via service-to-users examples, i.e. to see how a translated talk helps users, say in an old folks home, lab or at a rally. I hope/think the belonging motivation is in minority; most of us volunteer to help spread some ideas and improve our lingual traits. I suspect that in Chinese and other big tongues is the English learning aspect [inclusive fitness] more pronounced. This is why the system fails the novices... the best motivation for a generic new translator is to see the work has outcomes... it is published asap. However his/her subtitles are mostly published with some 3+ months delay. Bad system attributes: TED subtitling sports 2 bad system attributes> a] subtitles ownership exclusivity and b] rare LCs approvals. These are getting in the way of serving the users aka it is not working. Analysis: a] Sometimes anarchy is better than real socialism. Wikipedia is an anarchy of a benevolent community [all share the right of post edition] with little space for defence or paranoia [oh but it is not authorised oh ]. It gets translated by anybody and anybody can help to mend/submit better translations. It is OK to own a translation and be a proud provider but both systems should and can coexist. Say if subtitles get stuck in limbo for more than 5 months, it will get published automatically and then anybody can amend it... so say if you hold the ownership and value it over spreading of the idea, you would be lobbying at your community to barter for an early [less than 5 months] approval so that no co-translators get mentioned. [PS It is also sad that after I submit subtitles I can not improve them. When solutions (as they do) occur after a sound sleep, it is too late since the poor subtitles are locked, queuing in the review limbo... ] b] LCs: These valued multilinguals are translators who would not approve of subtitles with a mistake. They mostly do not 'coordinate' but act as the final controllers who have just time to dis/approve of reviewed subtitles. Since they do not use robots or automation and life happens to them, they do not volunteer with a planned/ideal focus and cause a 'grammar Nazi' situation: many subtitles are not published withing a novice retention span and so the communication flow suffers. Communication between the talk author, ted staff, translators and public/users. Novice retention is hard. Let's not pretend we have a know how. No research, no trials = no data...= we do not know! Same forces that say LCs have to write essays before their membership is agreed would say they know best how to retain novices. Bullshit. Allow trials. ...risk mistakes and try it on with parallel testing [pilot] approach. Solution ideas aka Pilots I would try -If a subtitle gets unreviewed [stuck] for 4 months it gets published as well as unlocked for amendments by anybody. [benevolent users] A communication business where the users can return. -A 2 speed approach> I would let another srt tools [not just amara] such as dotsub to be engaged in the endangered tongues subtitling [at least as a pilot]. In this 2nd pipeline subtitles ownership would be more collective, as it would not go through the LCs reviews/approvals but will be directly available at TED.com as a subtitle source [say community vs curated source] until it gets approved i.e. the curated/approved srt is available. Embrace the community, all of us, not just the curated club. So that fewer people would feel it [TED translating] amounts for nothing. The benefit of the approach is not only in higher publishing, but also in elimination of amara monopoly. Amara for TED has some UCD - issues. Its workflow design before actual subtitling is unintuitive, not user centered/friendly. All takes learning but Amara for TED should not be the only way on how to produce subtitles. You may assume the 2 pipeline system means worse quality, but we really do not know; we know it means quicker publication and so many more translations. -Another pilot is an AI [Deep Learning] mutually beneficial engagement. Google translate AI could be helped via LCs tribe cooperation. Feed it [let it learn] the review iterations [revisions history] of all translations to produce patterns for deciding upon where an initial translation [locked for review] is rather bad or rather good. Marking translation quality is another job but one with a result in an aide. I think Google would/should happily agree. Overworked LCs should feel some support and such a pattern-recognition aide should become a tool for quick approval decisions. [had the LCs be still needed] -Also: Had I spoken at a conference I would have been grateful for any volunteered subtitles ...am I nudged to help/pay the translators or reviewers? Many speak, milestone ticked out but many speak and their idea, their baby lives out there competing for attention. Mistakes as nudges: We do not know what a mistake in subtitles does/means... say one sees it and... there are these outcomes : - ..one can fix it, (in 2 pipelines system they thus become a co-translator) or - ..they curse the ted and the translator who dishonoured their best tongue and write an irate comment under the video, then someone replies or mends it or - ..they do nothing but their subconscious would later avoid further video sources with a poor translation quality or - ..they start to pay more attention to catch more mistakes [viewer's engagement] or - ..they are not sure so they check other sources [autonomous learning] being happy to be able to see a mistake (in 2 pipelines system they are nudged to become a co-translator) Some idea inspirations> https://seths.blog/2018/11/quality-and-effort/ https://translations.ted.com/Quick_start_guide_for_Language_Coordinators#How_to_post-edit_a_published_task https://www.youtube.com/watch?v=ZI96A7sYnLY
0 Comments
Leave a Reply. |
Authorbiodive.weebly.com Archives
August 2021
Categories |