Linguistic Team International Forum

Well, this problem is not yet solved, at least in my mind. The original location is where you get the video from. And the repository location is where people watch the video with subtitles.

The way we have been approaching this until now is to create a separate repo location in dotsub. It solves the problem, but I have been questioning this approach for some time now, because it creates more administrative work.

Another possible approach is to use the original location as the repo location, which would create minimum administrative work. And for situations where the original location is in a player that doesn't support subtitles (vimeo now does support subs, btw), we can create a youtube repo location in the LTI channel, since youtube seems to be handling subtitles for viewing much better than dotsub.

One of the core goals of LTI is to strive to massively increase global distribution of all of these materials, especially those that emerge in languages other than English, and we certainly don't want to inadvertently limit or dictate 'how' people can use the resources that come out of the hard work of the transcribers/translators. So, the approach I originally built into LTI's system is intended to provide the possibility for as many potential real-world "use cases" as possible. In other words, our intent is to provide everyone with the ability to be as creative as possible in spreading awareness as far and wide as possible. Simple, no?

The original location is always provided for those who may want to download the original video and add their language subtitles to it on a DVD (or multiple subs, if ambitious enough), perhaps for mass duplication to give away to friends &/or strangers, or perhaps to have a 'hard copy' for use in RBE parties (inviting a bunch of people over to your home to watch some good info).

On the other end of this is the intent to eventually make all completed transcriptions & translations available to the world in as many usable formats as possible, so that anyone will be able to pass around SRT (videos-only), text, PDF or even html versions of any given translation, to be used for spreading the info however they see fit.

I personally don't worry about how much administration it takes to accomplish these lofty aspirations. If/when some part of it proves to be too difficult for some reason, that simply means it's time to turn more of our attention toward developing automation for as much of it as is possible. That's why there's a continual need for developers to come join us to help simplify the manual tasks.


As for the public repository locations (where people can watch the video with completed subs), where the original is unable to support subs, we will simply support the subtitles globally within the LTI YouTube channel. In almost all cases (there are some that should never go up into YouTube), each language team should already be supporting their own subs within their own LTI-specific YouTube channel, with everything (titles, descriptions, 'about' page, etc.) written in their native language to attract people who do not understand English.

Lizzies words and way of thinking seem pretty clear to me. I don't find other ways to put this anymore, except metaphors and analogies with systems that have nothing to do with this - not the best way but...

When you have a cancer and you wish to make a diet towards the cure, all sources seem to agree: If there is a food that doesn't help cure it, even that it is not a cancer cause, it is just getting in the way of it. It is a waste of energy of your body in its goal to get better.

In the human brain, there is only one neuron between the olfactory system and the nervose centers, because in this case, it is an evolution from when we were mice or before. So, the other later neuronal pathways had to go around and add turns and twists so that they could connect their systems without compromising this olfactory one. Pretty expensive system but still made its way until today. After all, we aren't mice but we still use our noses. We can not simply rebuilt it from scratch, for obvious reasons. But in the LTI system, we can do whatever we want, because it is digital and in online platforms. We already have one pathway between the public 'out there' and the LTI, if we use the original locations of TZM and TVP to provide our subs, for watching or downloading.

Providing the subs in the original locations in youtube, for both watching and download, and safeguard them in the LTI server, seems to accomplish our goals pretty well and easy. If any specific situation doesn't allow for that, we might have to spend some more energy on that case. A new type of smell, what the heck!! This is getting fun! Aren't efficiency and abundance the understandings we are translating about as taking us to situations that make human behavior change? Why are we creating more bank accounts (accounts called LTI to represent the value of what we translate, others to have a back up in gold) when what we have is a problem of distribution of goods - goods not reaching everyone? That is the goal of what we are doing, no? Make RBE contents available to as many people as possible? I don't see most of the tasks we are engaged at as being focused on that goal. We were a lot of time trying to create an LTI 'image' to something we were not doing. If people still don't know where is the wiki repository or how to find a subtitle there, after all this time,maybe we could change something? Not that we are stuck to a million years evolved rhinencephalon, are we?

The purpose of the Repository location is to have a place where to put all finished translations of a text/video for the world to watch.

Technology is changing rapidly. I think that when the Repository was first set up in dotsub, youtube didn't support or was very limited in its support for subtitles. Now this is different and from my perspective youtube handles the presentation of subs much better.

The Repository location's other function is to be included in a planned Public Portal website that would present all translated materials to the world in a much better way. There is a difference between dotsub and youtube, since they 'connect' with the website in different ways. One has certain advantages in some ways, the other in others.

So, these are the two functions of the repo location. The two platforms (dotsub and youtube) handle these functions differently. I don't think it would be beneficial to go into details here about exact technical features. This is just an overview.

There are two different needs here, which is the thing that seems to be at the core of most confusion surrounding this question.

The ever-present need is to have all of these transcriptions & translations easily searchable and available to the public within the currently most-popular website (YouTube). The second need is for a consistent source feed for the LTI public portal, which is designed to bring all relevant RBE materials (not including websites) together for much easier learning by the public. Both dotSUB & YouTube are continuously evolving and it is not yet known which will provide the maximum flexibility & stability for the public portal's extensive public display of these materials. Initial testing places dotSUB in a huge lead over YouTube.

Of course, there is much more to the details surrounding every aspect of this that you could question, but it all becomes obsolete if/when we are able to automate most, if not all, of these tasks that are frequently brought into question. The obvious need to automate various tasks provides evidence of positive motion & movement, not negative ramifications of chores that no one wants to do. If that were the case, an analogy would be that the best way to handle higher demand for automobiles would be to produce fewer models and lay off more workers to keep 'the system' simple (don't allow need to get in the way of our comfort). This is why I said earlier that the various administration tasks that reveal themselves to us should be viewed as little more than 'the system requesting automation' from us, rather than viewing it as any kind of drag on our resources, excessive stress or any other negative connotation. How you think about it makes all the difference. I see massive potential for advancing our global distribution abilities, but not if we focus on keeping things simple and easy for humans to manage. Think 'automate' and the questions change dramatically.

After yesterday's meeting, it was agreed upon that supporting the LTI youtube and dotsub repository isn't serving us much right now. What is important is to have the finished subtitles in the originator's location. So for now we can stop uploading subtitle files to LTI youtube and dotsub repository (unless it is needed for a specific case) and concentrate only on the originator's location. This should ease the handling of administrative tasks such as ToDo's.

I find it also worth trying Amara.org, right on the new Zday videos, as a platform that:
- is very simple to create working locations;
- easy to transcribe and synchronize (timeshift) where even the "text adding and string structure" can be done by one person and the "synchronization" by another, and the "review" of the text even by another, as they are split in 3 stages there; but we can do them all by the same person too.
- we can click to "follow" one language or all of them, with email notifications when anyone makes changes;
- when anyone changes a translation it doesn't substitute it but actually creates a new version (in the new interface at least - beta);
- Another advantage is that it has comments in each language section for the video.
- The video is the youtube location, and when the amara location is added in the video description, people in one location can find the other easily and jump between them (viewing/repository <=> working location).
- Our forum.linguisticteam.org will be in the video description and so in the working location too.

I am testing with this amara location for the latest TVP Seminar, for a reference if anyone wishes to see how it works (or not) for the features we need: The Venus Project Seminar (9 March 2014) with subtitles | Amara

hugs : )

@Di Anna:
@Suze:
@FrankySnatra:
besides the others already in this thread,
Would be nice to know your thoughts on this.

EDIT (again):as we are talking about transcriptions, a valuable info I am confirming still today is that: to correct the string structure of a lousy made transcription is a pain in the ass, no matter what platform we use, and easily avoidable with abundant information and initiative from the most experienced in the English team to grab the projects right on the transcription stage.

Are these Amara locations being shut down once the English transcript is ready for final review? If not, it will encourage people to keep working on it there while we support the remainder of the project in dotSUB, encouraging senseless duplication of translation efforts.

No. There is no "shut down" or ways to limit access nor even deleting subtitles in Amara. There is no need for it either. Only one person can be editing one language location at the time in Amara and each new editing is saved as a new version. The focus has to become on doing and promoting what we do and what we intent / find important. There would be no efforts focused on processes of limiting but only on processes of doing and promoting/sharing what we are doing. If we use amara and dotsub at the same time, we are creating duplication, or developing extra efforts. So, when I am suggesting this, all I am visualizing in terms of actions to be taken are: use the amara platform to create locations, transcribe, proofread, translate and share information, on them and about them in this forum and in the facebooks and what else. So, only actions towards doing and promoting.

I think what each one of us can do is try out amara and then share their observations in a dedicated thread. Some time ago, when TED's translation project switched from dotsub to amara, I went and investigated that platform (it was called Universal Subtitles then). What I discovered were some strong advantages over dotsub, but also nasty bugs that weren't being fixed and probably overwhelmed developers, who that weren't able to give answers to people's questions. It would be nice to see if the platform has matured since then.

This is my proposal - everyone of us goes and checks out amara, then we share our findings in a dedicated thread. We would be able to better arrive at a decision that way.