| Pages in topic: < [1 2 3 4 5 6] > | Poll: Do you think DeepL is something for human translators to worry about? Thread poster: ProZ.com Staff
|
|---|
Tom in London
United Kingdom
Local time: 05:08
Member (2008)
Italian to English
| Backwards development | Apr 20, 2022 |
neilmac wrote:
....... as time goes by and the technology develops.....
You're assuming that the technology always gets better. Actually it gets more profitable for vendor - which may mean that it actually gets worse. Look at what Microsoft has done to its Office suite over the past 10 years or so.
| | | | Lieven Malaise
Belgium
Local time: 06:08
Member (2020)
French to Dutch
+ ...
| Contracts are MT material. | Apr 28, 2022 |
Rodrigo Bovino wrote:
You can't translate a contract or technical stuff with DeepL without MTPE and research.
Cheers
It's not about technical or non-technical, it's about the amount of context. If you have technical descriptions with sentences consisting of only a few words, MT is no good. But if you have to translate even really technical stuff that's poured into whole sentences and paragraphs, MT is perfectly fine as a basis for fast translation.
Also for contracts: I would even say contracts are made for MT, probably because the internet is full of them, so the MT memories are constantly fed with them. It's amazing how fast you can translate a contract nowadays compared to the old days, where you had to work your way through extremely long, strangely formulated sentences. Now it's already done for you for at least 75%.
EDIT: I see you're mentioning the need of MTPE. Of course there's always an amount of editing involved, but the truth is MTPE is a lot faster than normal translation, and that's what it's about.
[Edited at 2022-04-28 08:49 GMT]
[Edited at 2022-04-28 08:51 GMT]
[Edited at 2022-04-28 08:51 GMT]
| | | | Dominic D.
Local time: 06:08
Serbian to Russian
+ ...
Yes, contracts, GTSc, GDPR, EU texts, anything legal really...
| | | | jyuan_us 
United States
Local time: 00:08
Member (2005)
English to Chinese
+ ...
| MT can be of zero help in some topic areas | Apr 28, 2022 |
I just finished translating a few pages on the website of a company, and it turned out MT didn't increase my translation speed at all.
[Edited at 2022-04-28 09:38 GMT]
| | |
|
|
|
| Language pairs and law systems | Apr 28, 2022 |
Lieven Malaise wrote:
It's not about technical or non-technical, it's about the amount of context. If you have technical descriptions with sentences consisting of only a few words, MT is no good. But if you have to translate even really technical stuff that's poured into whole sentences and paragraphs, MT is perfectly fine as a basis for fast translation.
Also for contracts: I would even say contracts are made for MT, probably because the internet is full of them, so the MT memories are constantly fed with them. It's amazing how fast you can translate a contract nowadays compared to the old days, where you had to work your way through extremely long, strangely formulated sentences.
I guess this says something about how continental Europe does things; if you don't have to build insights into the intricacies of a completely different law system and come up with ways for connecting them to each other (which includes constructing terms that don't occur naturally in the target language but must nonetheless be understandable), then I guess you'll be able to delegate much of your work to a machine. In the English-speaking world, contracts are huge and often extravagantly worded, and their wordings vary a lot because clients pay lawyers to build custom contract templates for them. In my language space, the only good examples of legal translation that I know of are found in my personal TMs. And technical translations are all about having a strong intuition about how engineers write, which is something MT can't have. Just check out KudoZ questions (I'm only talking about my language pair). I once edited the translation of a Euronorm, where most of the translators just copied and pasted the Russian translation of that Euronorm (even where the English text said something completely different). The Russian text is found in an official document implemented by the government of Belarus, but the translation there is so horrible that I believe it's completely unusable (don't know if other Russian-language Euronorms they have are any better, but that one was an abomination). So there, like I already said in other posts, maybe there are language pairs that have a good potential for machine translation. Mine is not one of them, and most of the texts I translate are not machine-translatable. For some projects, the only thing that helps is professional literature, even dictionaries stop offering meaningful inputs
| | | | Tom in London
United Kingdom
Local time: 05:08
Member (2008)
Italian to English
QUESTION: Do you think DeepL is something for human translators to worry about?
No. Global warming is is something for human translators to worry about. Stupid politicians trying to start World War Three is something for human translators to worry about. Let's try to keep a sense of proportion.
| | | | Lieven Malaise
Belgium
Local time: 06:08
Member (2020)
French to Dutch
+ ...
| But often a big help. | Apr 28, 2022 |
jyuan_us wrote:
I just finished translating a few pages on the website of a company, and it turned out MT didn't increase my translation speed at all.
[Edited at 2022-04-28 09:38 GMT]
MTPE isn't suitable for everything, but in many cases it is a big help. Try identifying the ones that are a big help.
| | | | Lieven Malaise
Belgium
Local time: 06:08
Member (2020)
French to Dutch
+ ...
Denis Fesik wrote:
I guess this says something about how continental Europe does things; if you don't have to build insights into the intricacies of a completely different law system and come up with ways for connecting them to each other (which includes constructing terms that don't occur naturally in the target language but must nonetheless be understandable), then I guess you'll be able to delegate much of your work to a machine. In the English-speaking world, contracts are huge and often extravagantly worded, and their wordings vary a lot because clients pay lawyers to build custom contract templates for them. In my language space, the only good examples of legal translation that I know of are found in my personal TMs. And technical translations are all about having a strong intuition about how engineers write, which is something MT can't have. Just check out KudoZ questions (I'm only talking about my language pair). I once edited the translation of a Euronorm, where most of the translators just copied and pasted the Russian translation of that Euronorm (even where the English text said something completely different). The Russian text is found in an official document implemented by the government of Belarus, but the translation there is so horrible that I believe it's completely unusable (don't know if other Russian-language Euronorms they have are any better, but that one was an abomination). So there, like I already said in other posts, maybe there are language pairs that have a good potential for machine translation. Mine is not one of them, and most of the texts I translate are not machine-translatable. For some projects, the only thing that helps is professional literature, even dictionaries stop offering meaningful inputs
Good point. I always tend to think my translation experience is everyone's translation experience, but of course much will also depend of your language combination(s) when it comes to machine translation.
| | |
|
|
|
Daryo
United Kingdom
Local time: 05:08
Serbian to English
+ ...
| If the contract is one | Apr 28, 2022 |
Lieven Malaise wrote: Rodrigo Bovino wrote:
You can't translate a contract or technical stuff with DeepL without MTPE and research.
Cheers It's not about technical or non-technical, it's about the amount of context. If you have technical descriptions with sentences consisting of only a few words, MT is no good. But if you have to translate even really technical stuff that's poured into whole sentences and paragraphs, MT is perfectly fine as a basis for fast translation. Also for contracts: I would even say contracts are made for MT, probably because the internet is full of them, so the MT memories are constantly fed with them. It's amazing how fast you can translate a contract nowadays compared to the old days, where you had to work your way through extremely long, strangely formulated sentences. Now it's already done for you for at least 75%. EDIT: I see you're mentioning the need of MTPE. Of course there's always an amount of editing involved, but the truth is MTPE is a lot faster than normal translation, and that's what it's about. [Edited at 2022-04-28 08:49 GMT] [Edited at 2022-04-28 08:51 GMT] [Edited at 2022-04-28 08:51 GMT]
If the contract to be translated is one involving YOUR OWN MONEY I would strongly recommend you to make 101% sure that no MT was allowed anywhere near it at any time.
I know of cases where just few wrong words in a contract have sunk not so small companies.
Using MT for contract is asking for troubles, about as safe as crossing a road blindfolded.
| | | | Lieven Malaise
Belgium
Local time: 06:08
Member (2020)
French to Dutch
+ ...
| You have to use it correctly. | Apr 29, 2022 |
Daryo wrote:
I know of cases where just few wrong words in a contract have sunk not so small companies.
Using MT for contract is asking for troubles, about as safe as crossing a road blindfolded.
The whole point of machine translation post-editing is to identify and correct those 'few wrong words' in otherwise mostly correct sentences. Nobody is saying you should just accept the automatically generated translation. Machine translation is nothing more than a basis to work upon. But an increasingly good basis that allows you to work faster.
| | | | Daryo
United Kingdom
Local time: 05:08
Serbian to English
+ ...
| Although I would agree about fast food | Apr 30, 2022 |
Andreas Baranowski wrote:
Machine translation, whether DeepL or any other brand, will ultimately do what fast food did to restaurant cooking or digital photography to bromide photography. Although vastly inferior, I predict it will find acceptance on the merits of cost performance and acceptable fulfillment features – acceptable in the opinion of users. In the same way as fast food is a bad joke compared with the skills of a trained chef and digital will never deliver the nuance of bromide, machine translation will find its place and ultimately acquire an overwhelming market share. It’s just a question of time.
Although I would agree about fast food, when it comes to digital photography it's not so clear cut. The improvements in digital cameras and screen quality (resolution, colour reproduction etc ...) are making digital photography better and better, and comparable in quality to the best printed photography made from roll films. Yes, old style photography has its charm, but the digital photography of today is not the grainy crap from 10-20 years ago.
| | | | Tom in London
United Kingdom
Local time: 05:08
Member (2008)
Italian to English
| There speaks a man who has never heard of... | May 1, 2022 |
Daryo wrote:
Yes, old style photography has its charm, but the digital photography of today is not the grainy crap from 10-20 years ago.
...Don McCullin
...Henri Cartier Bresson
...Robert Capa
...Irving Penn
...Robert Doisneau
...Robert Mappelthorpe
...Bill Brandt
...William Klein
...Julius Schulman
...Gabriele Basilico
...etc.
https://www.lamag.com/longform/a-shot-in-the-dark/
[Edited at 2022-05-01 18:54 GMT]
| | |
|
|
|
Daryo
United Kingdom
Local time: 05:08
Serbian to English
+ ...
| As it happens | May 2, 2022 |
Tom in London wrote: Daryo wrote:
Yes, old style photography has its charm, but the digital photography of today is not the grainy crap from 10-20 years ago. ...Don McCullin ...Henri Cartier Bresson ...Robert Capa ...Irving Penn ...Robert Doisneau ...Robert Mappelthorpe ...Bill Brandt ...William Klein ...Julius Schulman ...Gabriele Basilico ...etc. https://www.lamag.com/longform/a-shot-in-the-dark/[Edited at 2022-05-01 18:54 GMT]
As it happens, here speaks a man who do know a thing or two about photography, and about most of them. And I'm sure many of them would not be adverse to using the best digital cameras of today.
But, to go back to the main point of this thread, it's not that it makes MT any more suited for a totally different type of tasks.
| | | | | Two kinds of professions | Jul 4, 2022 |
There are two kinds of professions:
For actors/actresses, fiction writers, artists, etc., they either become successful and earn huge money, or struggle at the bottom trying make ends meet. There's no room for them to learn and grow while being able to support themselves from income from the job.
For doctors, engineers, etc., they start at the bottom where they can still support themselves and gradually grow and most of them eventually climb to the middle where they can... See more There are two kinds of professions:
For actors/actresses, fiction writers, artists, etc., they either become successful and earn huge money, or struggle at the bottom trying make ends meet. There's no room for them to learn and grow while being able to support themselves from income from the job.
For doctors, engineers, etc., they start at the bottom where they can still support themselves and gradually grow and most of them eventually climb to the middle where they can live a comfortable life. A few reach the top and become famous. These professions allow people to learn and grow while working on the job.
What DeepL does is changing the translation profession in a way that deprives inexperienced translators of opportunities to spend many years on the job to learn and grow while being able to support themselves. And even if you eventually become a very good translator thanks to the years of financial support from your parents, the reward is not as thrilling as those enjoyed by actors, writers, artists, doctors, engineers. Even though DeepL may take years, or maybe never, to be able to compete with the best human translators, they have successfully change the industry forever. We cannot pretend that nothing has changed and that we can carry on the business as usual.
In fact, I think the industry is fucked, unless your idea of translation industry includes only those limited number of top translators, excluding all those students, teachers, young beginners. ▲ Collapse
| | | | | Yes, it's a big threat | Feb 28, 2023 |
Machine translation is coming strong into our industry and an increasing number of translation jobs are being transformed into machine translation post-editing (MTPE) jobs, often carried out on the basis of a machine translation produced by DeepL or a similar technology.
This comes together with price cuttings. People working for agencies are already being forced to accept considerable discounts (30-50%) for MTPE jobs even if no time savings can be seen if done properly. ("Forced" i... See more Machine translation is coming strong into our industry and an increasing number of translation jobs are being transformed into machine translation post-editing (MTPE) jobs, often carried out on the basis of a machine translation produced by DeepL or a similar technology.
This comes together with price cuttings. People working for agencies are already being forced to accept considerable discounts (30-50%) for MTPE jobs even if no time savings can be seen if done properly. ("Forced" in the sense that they'll lose work and income if they don't accept.)
The result of this development will be that traditional translators will be pushed out of the market (unless they accept producing low quality and work at lower incomes) and increasingly replaced by less competent workers who are willing to accept lower rates.
I would recommend all translators to prepare themselves for this development and start developing alternative sources of revenue.
[Edited at 2023-02-28 21:35 GMT] ▲ Collapse
| | | | | Pages in topic: < [1 2 3 4 5 6] > | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Poll: Do you think DeepL is something for human translators to worry about? | 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 » |
| | Wordfast Pro | Translation Memory Software for Any Platform
Exclusive discount for ProZ.com users!
Save over 13% when purchasing Wordfast Pro through ProZ.com. Wordfast is the world's #1 provider of platform-independent Translation Memory software. Consistently ranked the most user-friendly and highest value
Buy now! » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
