[Viki Community Team] French Machine Translations

[christina_]

In my experience, the CM problems, the irresponsible subber/segger/designer/coffee maker problems are 95% due to the system in place, which values quantity over quality. In my opinion, that is.

• Supposedly, CMs are hoarding projects and never finish them because they are irresponsible people. Also, they have no way to bootleg subs to TV channels nor they just want to bypass regional and other restrictions by becoming CMs, thinking that nobody is expecting subtitles after the first week of airing, anyway.

• Supposedly, english subbers can start working on live dramas immediately and they have plenty of skilled people to do the editing right after.

• Supposedly, after the english subs are done (and magically edited, polished and finalized), all other subs are good to go and should be complete really fast. Viewers are happy with a dialogue’s keywords, no need to spend an hour on an episode.

• Supposedly, CMs and mods are responsible for their team’s “product” and they have the means to evaluate it. Imagine what would happen if a bot appeared in the… Oh, wait…!

• Supposedly, Viki monitors quality and punishes any low standard subtitles.

• Supposedly, volunteers should contact a team before they can touch the sub/seg tools they find in various viki webpages.

• Supposedly, there are so many inactive/long-dead channels that nobody should have a problem completing their 500 subs, without destroying the hard work of others or begging to be in a team.

• Supposedly, Viki makes sure each of those 500, QC-worthy subs are real subs before they hand out QC statuses.

If you look at the problem this way, it makes sense to ask for ratings, public or not. However, all of my “supposedl-ies” have an explanation and it is 95% Viki’s poorly implemented procedures and only 5% contributor dramaqueening.

If those procedures were re-designed, the contributing community would most probably self-regulate, without ratings other than some fun motivational ones they seem to enjoy.

Also towards the friend who did a poor job or someone you didn’t really like to work with but did a good job?

I agree. Especially after having read some contributors’ input over the years. * rolleye *

[christina_]

but it will be nice if it takes into consideration how the word might affect the other language (

To do that, they would need to know at least the basics of the other language.
Also, they would need to sacrifice accuracy in their own set of subs while only guessing the results for other languages, that are not necessarily latin.

We can correct the flaws it has,

We certainly can, wasting time to fix something we could have written ourselves, correctly, the first time. Also, correcting a bad translation is harder than producing the original one, because our mind gets briefly confused after seeing the familiar (but inaccurate) terms and it has to reprocess the original meaning.

[irmar]

if someone can suggest a better objective solution to oppose the underhanded ways people get moderator positions or abuse their rights as CMs and moderators, be my guest.

We do talk about these things among ourselves, don’t we? We ask each other opinions about same language moderators. I know I’ve been asked and I have asked those I trust. We all have a personal list of “no-no” people.
Yes, it’s not a perfect system, and it’s still based very much on personal opinion, just like the public ratings would be. But instead of a simple number, when we explain why a person is not good, we usually give examples, tell the story of our past experience with that person - there’s more interaction.
When they do bad things we should report them and, most importantly, Viki should do something about it. Which it most often doesn’t.
When I reported to Viki an inactive CM for a channel, and asked them for the position, they first wrote to her, she said “yes, yes, I’m getting to it”, then another six months passed, with zero activity on the channel, then I wrote to Viki again, I got a reply saying “You two find a solution between you”!!! After another six months or so they finally gave me the channel. It took a tremendous amount of time. They don’t want to offend long-time volunteers who might leave Viki in anger.

Monitoring the community more closely would probably need more staff, and it would be costly for them. But it’s necessary. Now it’s only two-three people doing all the job. How can we expect them to take care the numerous instances?
A hoarder-checker would be a full-time job.

If the user Contributions page were better organized, it would be much easier to spot hoarders.
Put columns:

• added to the team date
• first worked on (with an option of Nul, if the person hasn’t touched the episode),
• last worked on,
• language (of the particular user)
• subtitle/segment completion %
• released to All Languages (yes-no)

And give the possibility to sort by each column (also a choice between decreasing and increasing order). So that we can sort the projects to see first the older ones and see their completion. Or the ones where the completion is less and see how many they are. Or the ones the user hasn’t even touched.
Of course moderators don’t always do subtitles nor always double as editors, so their sub count may be zero; however, they are responsible for the subbers doing the job, that’s why the completion percentage is important.
In case of CMs which language shall you choose? Let’s say the English. Although the CM is also responsible for all the other teams, at least if the English is complete AND released (thus presumably edited), you can say the most important part of her/his job is done.

[christina_]

The ‘explore’ page has a total of 723 shows (I remember they used to be thousands, but anyway). I guess most of the 723 have been there for years. It would shock me if Viki needed extra staff to check with a pace of 1 (one! ) show/day during its 5+ years of existence.

They don’t want to offend long-time volunteers who might leave Viki in anger.

Well…I wouldn’t bet my first-born on it.

Put columns:

Good idea. With minimal cost for Viki, I wonder why it isn’t in place already.

[lutra]

The ‘explore’ page has a total of 723 shows

Did you use any filter, as for me it says 2000!

[lutra]

as for me it says 2000!

But I just found out if I only chose tv content, it still says 2000!
If I chose movies it says 512, is it buggy, for “new” the oldest channels appear not at the last pages. What is it?

[christina_]

Did you use any filter, as for me it says 2000!

Not that I know of. I chose “all” in every category and when I saw 723 I double-checked! A new Viki user would pick those options, right? So they would only se 723! What did I filter out?

As I’ve been away for a while, I thought they removed the old fan channels and that’s why I can’t see thousands.Anyway, 2000, checking 1 per day, is still feasible to monitor, isn’t it?

[piranna]

I want to say thank you because trying to build something on my idea (I’d prefer to say “building on an idea”) or improving it is the basis of the creation process. And it’s beautiful and it’s what makes us advance in finding a solution.

Each scientist uses other findings to try and make new experiences. It’s a continuous process. We are bound to find weird results, to find contradictions to our original idea and to make mistakes that are contributions to create the system that would possibly work in our environment.
I know no child who knew how to walk alone without falling on the ground. We learn from each contribution, from each fall, from each step to be finally able to walk.

I found someone challenging ideas without feeling judged personally, without judging others personally while thinking. And I’m I’m grateful.
It’s the beginning of collaboration.

I don’t know what could be made to answer this question and that’s why we’re brainstorming, challenging ideas and hearing, receiving, analyzing ideas. I can auto-challenge my own ideas, but I know the limits of this. I need feedback, challenging, suggestions, it’s what thinking and creating is all about.
It evolutes.

It could be something AI could help volunteers with: building and suggesting many systems and we, humans, pick the system that is still ethically fine with us.

[idbarka]

I don’t know about the others, but I have 8 unfinished projests with very few segments simply because I did NSSA training…
Plus as there are few of us translating to my language, I am usuallly moderator and translator at the same time, have no language team members. I don’t need to communicate with CM except when I apply for the project.
What rate will I get then?

[bozoli]

I noticed that the suggestion tool (bot) in Spanish, certain words in English can affect the Spanish translation given as a suggestion. The subber here whom I know just clicks the FILL button, and won’t check the subtitle to see if it’s right or not, failed to see that the spanish translation was wrong in this sentence.

Come to think of it, are those two clicks worth the same on Viki as my personally crafted and sometimes several times edited sentence?? I don’t know why I didn’t think of that until now.

I think VikiBot suggestions should be valued differently, as half a contribution or something. Definitely not as whole sub. Maybe 5 fill-ins without any change should count as one standard manual contribution.

Ultimately GT might have a better performance than VikiBot. So, if we look down on people for using GT because they didn’t translate the sentence themselves, why are we valuing the application of the VikiBot suggestion equal as a human translation?

This is actually nonsense, when you think about it a bit!

if someone can suggest a better objective solution to oppose the underhanded ways people get moderator positions or abuse their rights as CMs and moderators, be my guest.

We do talk about these things among ourselves, don’t we?

Yes, we do…

But I feel the punishment system (VikiPolice) is completely ineffective. I love transparency. What we have atm is a mini society which talks behind each others’ backs, sometimes with good intentions, sometimes not, whether it is to try to push the project through or nepotisticaly protect your buddy.

Your second part of the suggestion is more-less what I suggested as Volunteer Statistics. It’s objective and easily calculable. That is if Viki would allocate some of it’s server function for it. Whenever there is a marathon, we have to apply by stating the initial number of contributions. This looks to me as if they can’t do basic deduction B-A for their whole volunteer database.

irmar:

Put columns:

Good idea. With minimal cost for Viki, I wonder why it isn’t in place already.

Current contribution page is more faulty than correct. It should show the last time I contributed something to the project. But, instead, it shows the first time I contributed to it (so the time I started it). Next, every time the licence on a show changes, whether it gets revoked or it comes back (yes, that has also happened, that the licence came back), again it says I contributed to the project recently, at the time of the licence switch. So, I wouldn’t count that the platform used to make contribution pages (which is also the platform for Viki) is the best option for user statistics. That’s why, should they try to really make the Marketplace, when choosing the platform to build Marketplace with, they could also have in mind Volunteer Statistics.

christina_:
The ‘explore’ page has a total of 723 shows

Did you use any filter, as for me it says 2000!

Don’t the two of you live in different countries?

It says 2.000 titles in total for me as well, but the fact that this number is so round makes me think it’s a bit false. True number could be 1989 or 2056.

[dudie]

I don’t know about the others, but I have 8 unfinished projests with very few segments simply because I did NSSA training…
Plus as there are few of us translating to my language, I am usuallly moderator and translator at the same time, have no language team members. I don’t need to communicate with CM except when I apply for the project.
What rate will I get then?

Ohh don’t worry about that. I even have some projects I didn’t even work on because of a Viki bug Viki still didn’t fix because according to them it isn’t a bug…

[anna79_9]

Hello everyone, viki contacted some of us for participate in a videocall to share our suggestions and recommendations, to express ourselves. I really thank you all for your involvement. We can stop this strike (action) and move on to communication now.

If some of you want to participate in this videocall, do not hesitate to send me a message. It will be certainly on Friday 27th September at 11 AM PDT. (https://www.viki.com/users/anna79_9/overview)

Send me a message in the upcoming some hours, please.

[bozoli]

I’ve been thinking a bit yesterday night why do I find the idea of user grade so clear cut, while you and others find it troubling. Perhaps you had much more negative experiences on Viki than I did, so you have a feeling there is more trolls out there to harm us.

But, perhaps it is also that my job, my projects culminate with an article which is peer reviewed. For those who don’t know what I’m talking about, briefly:

1. you do research on a project
2. you write a scientific paper with evidence and conclusions
3. you submit it to a scientific journal
4. the journal sends it to other scientists in the world, usually in your field who return your article sometimes saying it’s not good enough for publishing at all, and sometimes they say you have to do more experiments to be able to make your conclusion.
5. you get peer-reviewers’ comments back, swear at them internally and then read them once again and try to improve your article
6. you submit the improved version to the same journal and cross your fingers it gets approved for publishing.

Ultimately you are thankful to your tough reviewers because your baby, your project of sometimes sever years, has resulted in something more advanced and substantial than it was when you first submitted it to the journal.

So, although you get angry and frustrated because of all the comments, ultimately you know it’s for your own good. This type of (mostly) honest assessment I had in mind when thinking about users grading other users.

Viki’s own reality is different I know this. But I would like to skip the talks behind ones’ backs and just have it somehow indicated next to the profile pick where one could conclude: “Okay, this user is a great translator, but lousy in communication with others. Let’s put her on this movie project which she could easily finish on her own. And maybe that on-going high-profile drama project is not the right place for her after all, since it contains a lot of team communication.”

Or: “This user seems to be great at teaching new volunteers and getting them started. Because it seems most of the newbies she worked with have continued translating/editing. So, it’s okay if she has a team with more than one newbie and that she finishes the project a bit later.”

These are the kinds of assessments I would like people to make. Not “Omg… she has 2.4/5 average. I’m not giving her the mod position.”

[bozoli]

Hi, dear.

Are you sure it’s an open call, that anyone interested could participate? I figure they sent PMs to those they wanted to talk to…

[anna79_9]

We can invite some people if we want, so… It’s an open call, but not 50 persons, of course. ^^

[bozoli]

Hahaha! I would rate crapily for the same reasons as yourself. Too many unifinshed projects. In my case I couldn’t blame anyone but myself, for giving up translating something in Croatian which has already been translated either to Serbian (latin), Bosnian, Serbo-Croatian etc. I could clean up my house though - ask the mods to revoke my Moderator positions on those projects. But even after this I would “score” low.

I also translate a lot on my own. Volunteers in my language are rare, while the demand is high. Your CM could rate your progress, for example. If it takes me more than a year to translate a movie (guilty of it personally, that’s why I mention it), then I am not a good moderator after all.

I am also one of those OL mods who is not afraid to communicate with an English team (Editors) or the CM when I find something fishy in the English language. Or in the segments. Good OLs will be able to spot blatant mistakes and will have the will to report them back to the CM, who will in turn then activate the right team members to fix the error.

[lutra]

I am also one of those OL mods who is not afraid to communicate with an English team (Editors) or the CM when I find something fishy in the English language.

Guilty as well! Sometimes I think it’s strange, because so many teams already finished translating before me, but most of the time we clear a misunderstanding or a fault. Even as a viewer I do that if I see a faulty sub, I will write to the team, editor … Sometimes I think they read my name like, NO not her again. Once someone made a subtitle about necklace making it 24 kg, instead of 24 karats.

Sometimes I think it is a loss we no longer can see the actual subtitler, because in earlier years I would have sent them a personal “Thank You”. Most of them were happy to receive a PM like that, since normally you do not get a lot of recognition by the audience. In these days you will not look at comments any way, if you can avoid it. And it was a way for me to “recruit” subbers later on. I would have noted their names and when I started a project I contacted them and asked if they wanted to join. But now, you can’t tell anymore, and we all know how well the volunteer-finder works for us.

Hahaha! I would rate crapily

I just realize today, why the grading topic gave me goosebumps … Remember the one-liner project of Viki. A sentence by sentence without context, the outcome was not too far from the bot-translations with the problems alike. The bot would need an AI next to him telling it, that’s an ahjumma talking to a little boy and next is the CEO shouting at the ahjumma. Thinking back about what happened here at Viki I suddenly feel like a halmae “I have lived too long …”

[yorunome]

If the rates are only accessible to gold QC (so we reduce the possibility of people who would created accounts to give good or bad rates) and it’s only positive rate and a report? So if the volunteer is bad (google trad, etc.) you flag him but if he was good you give him a good review, if he was correct (like a 6/10 for you), you don’t rate him.

We can have categories: communication, moderation, translation, edition, segmentation, etc.
We can have different flag too: use google trad, too many CM projects, too many mod projects, etc.

[yorunome]

we can remove the flag option, so it’s just posivite rate, if you don’t like the person, you won’t be able to show it there.

[amiga_tauri_259]

I think that transparency is the key. If everyone has to provide a justification for the grades they give, it will be difficult to grade someone on the basis of being friends or enemies. As @bozoli said: honest assessment. Only hard facts, facts that one can confirm.