| Pages in topic: < [1 2 3 4] | MTPE is a dirty word... but client will not pay the rates for full human translation. Thread poster: BabelOn-line
|
|---|
BabelOn-line
United Kingdom
Local time: 17:14
English to French
+ ...
TOPIC STARTER | Ok. Minus the architects then. | Aug 3, 2022 |
Tom in London wrote: BabelOn-line wrote:
...... you can't hire an architect .....out of the country. .
Actually, you can.
Ok. Minus the architects then. What about the other examples quoted?
| | | |
BabelOn-line wrote:
Doubt they would. This is something they would not sell onwards. More something they have to provide. Hence why they want just good enough quality.
I don't think it's achievable. You need a hefty investment for this and even with bare-bone MTPE, the expense might be too big and not of good enough quality. It is what it is. The volume is too big.
| | | |
BabelOn-line wrote:
Did you ask what this speciific project is about? No.
Did you ask to see a sample of the MT output to get an idea? No.
You said earlier: 'over 30 projects, all requiring different specialist knowledge or a sound understand of the topic at hand – and probably 30 different linguists.'
If you want to tell us more, please do.
BabelOn-line wrote:
Again, you could simply say "not for me", you are perfectly in your right to explain your dislike of MTPE, to tell that in your experience MTPE was awful for you - in some given circumstances, which many other linguists have done. Qualify things. Differentiate.
I didn't say that I dislike MTPE or that it was awful. That's your words, not mine. Please don't claim that I said something I didn't. I will repeat what I said: sometimes MT can save a lot of time, sometimes it saves nothing. I have also said that formulaic texts such as T&Cs can benefit. But if you choose to ignore what we write, only to deform it, we can keep saying the same things over and over again.
You also said: 'the jobs in question require a sound knowledge of each topics - not necessarily at PhD level, but still. This is why we can't merely rely on "generic MTPE linguists",' which is one of the reasons for my scepticism.
And you said: 'For the first two jobs we processed, our three specialist linguists created translations that were in my view far superior to the source (ambiguities were cleared, many mistakes corrected, style was more fluid, etc).' Indeed. MT doesn't do that.
Then you said: 'I can't help to notice that no one on this thread asked if the aim was to reduce the costs by 10%, 20%, 50% or 80%.'
But you had already said: 'they will not order unless we can drastically reduce the rate per word,' so we can infer that it isn't about 10% or 20%. 80% would be pure fantasy, so it must be somewhere in the middle. You had already provided the approximate information, which is presumably why nobody is asking.
In other words, you require specialist knowledge, so it is unlikely to be formulaic texts, and at the same time you want a significant reduction. It doesn't sound realistic, which is presumably why not many are interested.
| | | | Sadek_A 
Local time: 20:14
English to Arabic
+ ...
| My second post in the thread | Aug 3, 2022 |
I just want you to know that Machine Translation will keep chasing its tail once it reaches a certain % of accuracy. I'm thinking it has now, and my proof is you starting this thread.
I talked before about links being made between terms across the varied contexts to pick the correct translation. I think now the easy links have been connected.
What is left at the moment are the hard links, the ones that need not just language experience but also critical & analytical thi... See more I just want you to know that Machine Translation will keep chasing its tail once it reaches a certain % of accuracy. I'm thinking it has now, and my proof is you starting this thread.
I talked before about links being made between terms across the varied contexts to pick the correct translation. I think now the easy links have been connected.
What is left at the moment are the hard links, the ones that need not just language experience but also critical & analytical thinking.
Those will require what I call Lingual Algorithmians. Think of them as engineers of linguistic algorithms.
Sadly for your and other MT developers' intents and purposes, those LAs will NOT sell cheap. So, any developer who is serious will:
- Present specific MT linguistic problems in desired pair, as a test.
- Ask for and receive solutions from interested candidates.
- Hire the successful ones at a VERY rewarding monthly salary.
- Expect to have a near-perfect product across all domains, in no more than 5 years.
End of story.
The developer will cash out BIG TIME from the product. The lousy wannabe, amateur joiners of this profession will start flocking elsewhere. The gems of the profession will have the salaries they collected building the product, which they will need to ration or invest because Translation is already Bye Bye For Good by then. That way everyone will know where they stand at that given point of time.
But, now, we are all **cken absurdly aiming at each other with everybody missing every single time. That is what should be called Translation's Stupidity-Induced Recession. ▲ Collapse
| | |
|
|
|
| Just an example | Aug 3, 2022 |
BabelOn-line wrote:
If I could offer a job that paid $200 an hour, I'd be happy to. It means I would in turn be able to bill this to the client and scrape a margin on top. Or I would do it myself!
BabelOn-line wrote:
I can't tell how much productivity gains can be expected precisely because we would have to do a small dry run to find out.
$200 was just a random number. If the gains on MTPE were that good, we'd all be beating your door down. My point was simply, "What's it in for the translator?"
You should do the dry run so you can find out and communicate what this job has to offer the translators you want to work with. But IMO it's going to be hard to get away from the fact that translators want to translate and MTPE is not translation. I still think it would be best to start assembling and training a team of MTPE specialist or MTPE-positive translators to handle such projects in the future.
[Edited at 2022-08-03 19:43 GMT]
| | | | Edwin den Boer
Netherlands
Local time: 18:14
Member (2009)
English to Dutch
| Let them be more selective | Aug 7, 2022 |
Thomas T. Frost wrote:
I have a solution: don't translate those millions of words. Simple. End of story.
This is the best answer, but it doesn't need to be simple. In my experience, clients pay zero attention to selecting which documents need to be translated. There's no process to filter texts for localization, there's no job title for someone who does this, localization managers or product leads are busy already, and adding an extra step would increase the time-to-market.
But if they did pay attention, they might find for example that only help texts for regular users need to be translated, not the longer, more detailed documentation for system administrators, which would be full of untranslated English terms anyway. So you could get the same results for 1/3 of the cost, without lowering the quality of translations.
I'd even question whether all these source texts needed to be written. In IT, most of our translation efforts are wasted on documentation that wouldn't be needed if the user interface was more intuitive. The cost of translating documentation is never considered when adding a feature. Much of this documentation is even computer-generated. Then you get helpful explanations like "The mold truncation module is used to truncate the mold."
On the other hand, if they simply want to translate everything as a legal requirement, then you're overthinking this, and the client will need to find a less scrupulous agency.
| | | | | I USUALLY take MTPE | Aug 13, 2022 |
Yes, I do take MTPE. It´s discouraging, though.
But I need the money, and it helps.
So, dear agent owners, you can count on my humble service.
Dear colleagues, I wish you all success, especially getting not MTPE texts...
All the best,
Rosie
| | | | Sadek_A
Local time: 20:14
English to Arabic
+ ...
| Then, let them not pay, and instead take their chances with MT; that will be amusing. | Aug 13, 2022 |
I wish them success and good luck in their endeavors.
My analysis of GT's translation of my last post from this thread:
- [chasing its tail]: literal
- [I'm thinking it has now]: failed to correctly segment [I'm thinking] [it has now]; turned it into [I'm thinking of it now]
- [hard links]: mistranslated as [solid] instead of [difficult]
- LAs: mistranslated as [local agencies] instead of [Lingual Algorithmians] which was right in ... See more I wish them success and good luck in their endeavors.
My analysis of GT's translation of my last post from this thread:
- [chasing its tail]: literal
- [I'm thinking it has now]: failed to correctly segment [I'm thinking] [it has now]; turned it into [I'm thinking of it now]
- [hard links]: mistranslated as [solid] instead of [difficult]
- LAs: mistranslated as [local agencies] instead of [Lingual Algorithmians] which was right in the previous paragraph.
- [will: - Present/- Ask/- Hire/- Expect]: failed to maintain connection between [will] and the verbs, turned the verbs from future into command instead.
- [cash out BIG TIME]: mistranslated into [spend a lot of time] instead of [collect a lot of money]
- [gems]: literal
- [collected building the product]: mistranslated into [collected to build the product] instead of [collected from building the product]
- [Bye Bye For Good]: literal
- [**cken]: couldn't understand what the word was, and just put the two asterisks in its place removing the remaining letters
- [aiming at]: mistranslated into [aiming to]
- [everybody missing]: mistranslated into [everybody went missing]
- [Translation's]: failed to recognize the significance behind capitalization, mistranslated into [translating's] instead of [Translation Industry's]
13 major errors in such an easy text; not because there are no qualified HR to build a better product, but because the aspiration is to build it while paying only pennies.
Was spotting those errors easy? No, it most certainly was not. ▲ Collapse
| | |
|
|
|
BabelOn-line
United Kingdom
Local time: 17:14
English to French
+ ...
TOPIC STARTER | Closing note | Aug 15, 2022 |
Rosa Coro Antich wrote:
Yes, I do take MTPE. It´s discouraging, though.
But I need the money, and it helps.
Thanks for your post, Rosa.
I originated this post and received – amongst others – many messages bordering on abuse (or at least clearly implying "you try to exploit translators" or "you clearly do not understand the first thing about translation as you should know that MTPE does not work", etc).
It hurts. It hurts as nothing is further from the truth. We value our linguists, it would have taken a quick glance at our Proz Blueboard to find out which kind of agency we are.
MTPE is here to stay. I don't like MTPE but have to admit MT gets better all the time. Some translators went out of their ways to show me examples of poor MT output. Never said it was perfect. Never said it work for all topics, syntaxes, language combinations.
We humans have peaked in translation a few centuries ago, the human/machine gap will only narrow. All you can hope is that it is not going to narrow too quickly (and I am sure the guys at DeepL or Google are working hard at it) and eradicate the profession.
So I feel your reply is very much a rational one: let's take as many full human jobs as we can, but for some application, it will have to be MTPE, and those who need the money will take the jobs.
I am really sad as the whole point of my post was to find out why I could not engage rationally with the profession regarding new approaches - namely, how can we get the MTPE "good enough" level.
I have this fear: if the top end of the profession decides that MTPE is "a dirty word" and they do not want to have anything to do with it, I am 100% sure that some agencies won't have any moral issues hiring non-linguists / non mother tongue speakers to process MTPE jobs.
MTPE would irremediably become a cheap and dumbed down market. Once the word is out that you can get translations for a few cents a word, even if it is bad, there is no turning back. Good luck trying to sell human expertise in such a market.
This is not was I was advocating at all. I was merely trying to see if there was a way of rethinking translation approaches. Valuing human expert input while leveraging MT.
I will no longer read entries to this post as this is becoming pointless and I feel like I am somehow the enemy here. I will however try to find other ways to engage with the linguists amongst us who think that feel it is probably better to try and tame the MTPE dragon rather than become its victim.
| | | | | I'm not against MTPE, personally | Aug 16, 2022 |
It's here to stay, so I agree with you that we - translators - should get together to offer our input, explaining how can be used better and in a more productive way. But the reward has to be there. It's normal that some of us will be against it and also quite vocal about it.
I know you won't be reading this, but I wish you good luck with your project regardless.
G
BabelOn-line wrote:
Thanks for your post, Rosa.
I originated this post and received – amongst others – many messages bordering on abuse (or at least clearly implying "you try to exploit translators" or "you clearly do not understand the first thing about translation as you should know that MTPE does not work", etc).
It hurts. It hurts as nothing is further from the truth. We value our linguists, it would have taken a quick glance at our Proz Blueboard to find out which kind of agency we are.
MTPE is here to stay. I don't like MTPE but have to admit MT gets better all the time. Some translators went out of their ways to show me examples of poor MT output. Never said it was perfect. Never said it work for all topics, syntaxes, language combinations.
We humans have peaked in translation a few centuries ago, the human/machine gap will only narrow. All you can hope is that it is not going to narrow too quickly (and I am sure the guys at DeepL or Google are working hard at it) and eradicate the profession.
So I feel your reply is very much a rational one: let's take as many full human jobs as we can, but for some application, it will have to be MTPE, and those who need the money will take the jobs.
I am really sad as the whole point of my post was to find out why I could not engage rationally with the profession regarding new approaches - namely, how can we get the MTPE "good enough" level.
I have this fear: if the top end of the profession decides that MTPE is "a dirty word" and they do not want to have anything to do with it, I am 100% sure that some agencies won't have any moral issues hiring non-linguists / non mother tongue speakers to process MTPE jobs.
MTPE would irremediably become a cheap and dumbed down market. Once the word is out that you can get translations for a few cents a word, even if it is bad, there is no turning back. Good luck trying to sell human expertise in such a market.
This is not was I was advocating at all. I was merely trying to see if there was a way of rethinking translation approaches. Valuing human expert input while leveraging MT.
I will no longer read entries to this post as this is becoming pointless and I feel like I am somehow the enemy here. I will however try to find other ways to engage with the linguists amongst us who think that feel it is probably better to try and tame the MTPE dragon rather than become its victim.
| | | | Sadek_A
Local time: 20:14
English to Arabic
+ ...
BabelOn-line wrote:
Once the word is out that you can get translations for a few cents a word, even if it is bad, there is no turning back.
And, once the word is out that: a defendant was MT-translated into admitting a crime he/she neither did nor admitted to in their source language; a wife was MT-translated into telling her foreigner-husband's family, in her letter to them, that she got pregnant by someone else other than their son; the fruit exporter is MT-translated into confessing presence of forbidden chemicals that are NOT there; etc; etc; etc; the U-turn will be crowded, congested, cramped, jammed, loaded, packed, swarming, overpeopled and full to bursting.
-----------------------------
BabelOn-line wrote:
Thanks for your post, Rosa.
Rosa and your good self, BabelOn-line, do collegially owe it to us to let us know how things will go between you two, MT-cooperation wise. Perhaps if you can make it, so can the rest of us!
BabelOn-line wrote:
I originated this post and received – amongst others – many messages bordering on abuse (or at least clearly implying "you try to exploit translators"
------------
1. lack of appetite for "giving it a go" equates to letting these dark times come without a plan B
2. engage with the linguists amongst us who think that feel it is probably better to try and tame the MTPE dragon rather than become its victim.
3. All you can hope is that it is not going to narrow too quickly (and I am sure the guys at DeepL or Google are working hard at it) and eradicate the profession.
Tell the truth: what do you call those 3 instances above, if not (emotional/psychological) blackmail?
Who is abusing whom here exactly?
BabelOn-line wrote:
It hurts. It hurts as nothing is further from the truth.
😭😭😭
BabelOn-line wrote:
MTPE is here to stay.
It should!
BabelOn-line wrote:
Some translators went out of their ways to show me examples of poor MT output. Never said it was perfect. Never said it work for all topics, syntaxes, language combinations.
No, that wasn't showing you. That was showing the developers. They need to know that they've been wasting their time and efforts, and that the triangle of success will not be complete without the third side: loads of money.
BabelOn-line wrote:
I am really sad as the whole point of my post was to find out why I could not engage rationally with the profession regarding new approaches - namely, how can we get the MTPE "good enough" level.
But, how could you ever, BabelOn-line? Engaging rationally with the profession requires you presenting a rational framework first, and you didn't present that nor are you willing to do so!
BabelOn-line wrote:
I have this fear: if the top end of the profession decides that MTPE is "a dirty word" and they do not want to have anything to do with it, I am 100% sure that some agencies won't have any moral issues hiring non-linguists / non mother tongue speakers to process MTPE jobs.
Hence, my Sadek_A wrote:
that will be amusing.
BabelOn-line wrote:
MTPE would irremediably become a cheap and dumbed down market.
What kind of a market is it now?
BabelOn-line wrote:
I will no longer read entries to this post as this is becoming pointless and I feel like I am somehow the enemy here.
No, don't say that, you're no enemy. Being so keen on having us develop the very product that will drive us out of business, with us getting pennies doing so doesn't make you the enemy. You're being hard on yourself!
Last but not least, I want everyone to understand this:
A perfect or near-perfect Machine Translation means no more:
- Translators
- Subtitlers
- Interpreters
- Dubbers
- Voice-Over Artists
- Tour Guides
- Language Teachers/Tutors
Because, voice-delivery-systems are already highly advanced now, and with linguistic-delivery-systems perfected, all those professions above are eliminated for good!
[Edited at 2022-08-16 18:08 GMT]
| | | | | Oversensitive? | Aug 17, 2022 |
BabelOn-line wrote:
I originated this post and received – amongst others – many messages bordering on abuse (or at least clearly implying "you try to exploit translators" or "you clearly do not understand the first thing about translation as you should know that MTPE does not work", etc).
It hurts. It hurts as nothing is further from the truth. We value our linguists, it would have taken a quick glance at our Proz Blueboard to find out which kind of agency we are..
It’s not uncommon in a discussion forum for people to have opinions that differ from your own 🤷♂️
| | |
|
|
|
Jeff Whittaker
United States
Local time: 12:14
Spanish to English
+ ...
| Rates for machine translation | Aug 17, 2022 |
There has to be some kind of concession from the client in exchange for a lower rate (ex: the result will be for in-house use only and not suitable for publication, we will have someone in-house re-write it, we understand that the syntax will not be perfect, etc.). If I have to provide the same quality, there is no reason for a discount because MT does not save very much time or mental effort for most projects. If you think it does, then you are missing something.
A State of Flux -... See more There has to be some kind of concession from the client in exchange for a lower rate (ex: the result will be for in-house use only and not suitable for publication, we will have someone in-house re-write it, we understand that the syntax will not be perfect, etc.). If I have to provide the same quality, there is no reason for a discount because MT does not save very much time or mental effort for most projects. If you think it does, then you are missing something.
A State of Flux - The Difficult (and Controversial) Task of Integrating Machine "Translation" Tools into Professional Translation:
https://www.languagecrawler.com/2021/04/a-state-of-flux-difficult-and.html ▲ Collapse
| | | | | Rates for machine translation | Aug 18, 2022 |
The rate has to be appealing. As it says in the article, if we have to work for a lot less, we will have to work a lot faster, and this will lead to missing errors, etc. In my opinion, MTPE will only be able to generate marginal savings if the results has to be near perfect or perfect. It seems to me that many clients don't understand this concept. They want you to work fast for a lot less and this is not conducive to an acceptable product, unless that's the client's decision. Agencies exploit C... See more The rate has to be appealing. As it says in the article, if we have to work for a lot less, we will have to work a lot faster, and this will lead to missing errors, etc. In my opinion, MTPE will only be able to generate marginal savings if the results has to be near perfect or perfect. It seems to me that many clients don't understand this concept. They want you to work fast for a lot less and this is not conducive to an acceptable product, unless that's the client's decision. Agencies exploit CAT Tools and MT/MTPE to be able to offer substantial discounts with the aim to attract more business, passing the financial pressure onto the translator. Unfortunately, a decent translation will always be expensive, because involves special skills and you can't really take big shortcuts. ▲ Collapse
| | | | | Pages in topic: < [1 2 3 4] | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » MTPE is a dirty word... but client will not pay the rates for full human translation. | Protemos translation business management system | Create your account in minutes, and start working! 3-month trial for agencies, and free for freelancers!
The system lets you keep client/vendor database, with contacts and rates, manage projects and assign jobs to vendors, issue invoices, track payments, store and manage project files, generate business reports on turnover profit per client/manager etc.
More info » |
| | TM-Town | Manage your TMs and Terms ... and boost your translation business
Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
