| Pages in topic: < [1 2] | editing a high quality machine translation Thread poster: Bernhard Sulzer
|
|---|
Mark Hamlen 
France
Local time: 17:09
French to English
+ ...
| I've decided to refuse all "Proofing" & "Editing" work for agencies | Jun 1, 2011 |
I've already been burned. Receiving utter rubbish for "proofing" that requires complete re-translation and the agency wants to pay a fraction of translation rates. Luckily I have enough translating work. I refuse to review work for agencies because of the few dishonest agencies. I will continue to do this for direct clients: reworking their documents written by non-natives and reviewing their in-house translations.
| | | | Bernhard Sulzer
United States
Local time: 11:09
English to German
+ ...
TOPIC STARTER | a new market for trashy translations | Jun 1, 2011 |
Neil Coffey wrote:
The developers of MT systems and the computational linguistics community more broadly are generally quite open about the fact that MT offers a poor quality translation for cases where an instant, cheap, poor quality translation is better than no translation at all. ...
So overall, I don't necessarily see MT as a threat to well-paid, high quality translation. I see it more as aiming at a different market... in principle.
[Edited at 2011-06-01 11:06 GMT]
Neil, thank you for your input.
I talked to a colleague and he confirmed the following:
there are agencies offering to their clients
1) machine translations only
2) machine translations + "light" edit
3) machine translations + "in-depth" edit
4) human translation
5) human translation + light edit
6) human translation + "in-depth" edit
The client decides what he/she wants and, accordingly, pays less or more.
Maybe some agencies really call it "high quality machine translation" to make it sound like it's pretty good. That of course would be deceptive.
And if it's just rubbish and the clients are okay with it, why are they okay with it?
Is there a market for it? There seems to be one. But how big is it? Who seriously wants to just put rubbish out as a translation of their website for example?
That's like selling bad cake (instructions) because it's cheaper than fresh cake (correct instructions)? Don't mind if you or your customers get violently ill/have an accident.
Where is the business ethic? I guess none is applied.
Maybe a disclaimer will read:
Warning! this is only a machine translation but it's a "high-quality" machine translation!
Warning! Use at you own risk. Target text does not necessarily convey the meaning of the source text
or:
Warning! This translation is cheap and of poor quality but it shows many words "of" the target language.
The question is: is it just a different, small market for some clueless people or will it hurt the real high-quality translation market?
I guess we will see.
Again though, the topic is really "editing a high quality MT" is not "editing" because there is no such thing as a "high quality MT" and I am afraid many times this just means the agency does not want to pay for a real translation.
Bernhard
[Edited at 2011-06-02 00:26 GMT]
[Edited at 2011-06-02 00:42 GMT]
| | | | Neil Coffey
United Kingdom
Local time: 16:09
French to English
+ ...
Bernhard Sulzer wrote:
Is there a market for it? There seems to be one. But how big is it? Who seriously wants to just put rubbish out as a translation of their website for example?
So I confess I don't understand the overall picture, but my theory is that the market concerns cases where in the past translation would never have been done "properly" but just passed to employees with a smattering of language knowledge to "give it their best go" and where good quality wasn't necessary (internal memos between subsidiaries, Chinese motherboard manuals...). With cheap on-line services and MT, I'm guessing companies are finding that they may as well off-load these "non quality-critical" jobs to a cheap service than occupy an employee's time with them when that employee could be doing something more useful to the company. So as I say, in this scenario, it's not that MT is "stealing" a genuine translation job as such, it's just more that it's creating a kind of "para-translation" market.
But I don't have any actual concrete figures on this: it's just a suspicion. If anybody does have any more concrete information, I'd love to be enlightened.
| | | | Neil Coffey
United Kingdom
Local time: 16:09
French to English
+ ...
| Assume innocent until proven guilty? | Jun 2, 2011 |
Bernhard Sulzer wrote:
Yes, that might well be the case. Elsewhere in the job post it says: "...Please only apply if your language capability is close to that of a native English speaker because you will be editing for correct English."
Wow, no kidding, "editing" for correct English. Because what you're going to be editing is certainly not correct English.
Calling this job an editing job is like saying give us your best "editing rate".
Why would we who have profiles here want to be associated with such business practices? It's like condoning them.
A priori, I think you may be being a bit harsh, though. I would read "high-quality" as a relativistic term: giving them the benefit of the doubt, I would suggest that they don't mean "high quality compared to a skilled human translator" but rather something like "high quality compared to the MT systems of the 1970s".
If edited machine translation is being passed off as a "real" high quality human translation and/or it is expected that the editors will be paid as though they were editing a quality human-produced text, then I would have similar concerns. But that's not a priori the case here. If managed and sold appropriately, corrected MT is a legitimate product, I think. It's not the same as a high quality human translation, the processes for handling it a different and it serves different purposes. And we face a challenge in ensuring that people are aware of that. But at the same time, we should be careful not to make unfounded allegations.
| | |
|
|
|
Bernhard Sulzer
United States
Local time: 11:09
English to German
+ ...
TOPIC STARTER | affordable and terrible translations | Jun 2, 2011 |
I think what you meant below is that many customers don't care how bad the translation is as long as it is cheap/affordable?
Samuel Murray wrote:
Don't make assumptions about what translation customers want. Many customers want an affordable product more than anything else (or is that also a sweeping assumption?).
I just don't want to be part of providing cheap and terrible translations or work for peanuts to fix up a MT. Wouldn't be good for my reputation (possibly) or wallet. And I don't condone the business practice of using bad translations in general, for whatever purpose. But maybe that's an extreme standpoint?
[Edited at 2011-06-02 03:16 GMT]
| | | | | A reasonable point | Jun 2, 2011 |
Bernhard Sulzer wrote:
I just don't want to be part of providing cheap and terrible translations or work for peanuts to fix up a MT. Wouldn't be good for my reputation (possibly) or wallet. And I don't condone the business practice of using bad translations in general, for whatever purpose. But maybe that's an extreme standpoint?
I think some people mix two very different things, and tend to think that machine translations go in the same boat as human translations. The fact is that we human translators (at least the professional, full-time translators) belong to a separate market.
Machine translation developers and companies using their products use machine translation to cover a market segment we do not work in: the low-cost, nobody-will-notice market. I do not work in that market --I am simply unable to worsen my translations intentionally--, so I cannot help them in their projects.
| | | | Bernhard Sulzer
United States
Local time: 11:09
English to German
+ ...
TOPIC STARTER | impact of "high-quality" machine translation editing on professional translator rates | Jun 3, 2011 |
Tomás Cano Binder, CT wrote:
I think some people mix two very different things, and tend to think that machine translations go in the same boat as human translations. The fact is that we human translators (at least the professional, full-time translators) belong to a separate market.
Machine translation developers and companies using their products use machine translation to cover a market segment we do not work in: the low-cost, nobody-will-notice market. I do not work in that market --I am simply unable to worsen my translations intentionally--, so I cannot help them in their projects.
Thank you, Tomás.
Getting back to my example I gave at the beginning of the thread, there it is suggested that 1) the MT is "high quality" - implying it is better than your "average" MT - and to top it off, 2) "real" human editors are hired to "edit" the translation.
So here the MT is entering our market segment if you will because it addresses professionals and invites them to take on an "editing" job which I believe is agency speak for "we want your best work for the best rate".
It's another way of getting professional work for cheap. That's not right and if "editors"/translators go for it, it might hurt "our" market segment or the individual translator because she/he is not going to get the job but the cheapo who accepts the terms will. Hope that good translators will stay away from it - that's one way to guarantee these things won't come back to bite us in the...
If good translators fall for it, it could change expectations of price and work volume etc. - once again!
That's why I say professional translators should speak out against it and/or not accept these jobs.
Nothing against CAT tools as a great technical aid, but agencies do and did use them to lower "professional" translator rates as much as possible - and what were we able to do about it, generally speaking?
The more we allow agencies to dictate how translations are done, the worse the result will be for translators overall. I am afraid that client expectations are being lowered - they'll be told well that's what you get for this price and it's a good enough standard.
Therefore I refuse to "edit" a machine translation. I would tell the agency that I would translate the source text because that's what I would end up doing anyway, even with an MT. And it will cost what a correct translation will cost.
I am not going to be duped into a difficult job that pays an unrealistic amount of money and I can only hope that's true for all professionals. But there are always the people just starting out in this business, what will they do - you know whatever job is better than no job.
If this is a project of some kind, as an earlier comment in this thread suggested above, to improve machine translations, then it should be stated as such.
Maybe it is possible for a machine to store millions of sentences in millions of different contexts and spit out a translation of a text that is already stored (ST and TT) but if the machine encounters just one word or phrase in a sentence in the ST that is different, how is it going to deal with it? And it will encounter many many things that aren't stored yet, I would say. It would need the human mind to understand its meaning within the context. Then it is still a long way to delivering a great/appropriate translation.
So what does a super MT actually do? It would have to think as we do. We are not there yet and hopefully will never be. Otherwise, we can pack in.
But to make us work for even less (do an "editing" job) because we got another "fantastic" tool, the "high-quality" machine translation, that is a danger on our horizon.
Bernhard
[Edited at 2011-06-03 23:28 GMT]
| | | | F Scott Ophof (X)
Belize
Local time: 09:09
Dutch to English
+ ...
| Might be better than you seem to think | Jun 7, 2011 |
Bernhard Sulzer wrote in his original posting:
"Company XXX plans to start a major program for editing high quality machine translation from several languages (European, South American, and Asian) into English.
A 4 or 5-year university degree is necessary to qualify. If you are a translator with native or near-native English language skill and are interested in participating in this long-term program, please e-mail your resume to ....."
It mentions "translators", not "editors". XXX does seem to want translators to do edit work, but recognizes the need for them to have a very good background in translating. So I'd say that XXX seems to at least have the right idea. Am I assuming too much when I say that I feel that a really good translator is also a really good editor and really good proofreader? (Oh, I do realize that proofing one's own work is almost useless within the tight time frames common these days, but still...)
I realize professional translation customers will always expect a "real" person to translate any text and a "real person" to do any additional editing, but it seems there is a lot of room for "using" editors and turning them into "translators" for the price of an "editor".
Is there really that much difference in the price of an editor versus that of a translator? Really, I hope not; the work of editors--and proofreaders--is very necessary. I don't have a 4 or 5-year university degree, else I'd ask you, Bernhard, to please pass me the full text of this job posting....
Somehow I have a feeling that the final text will not be a high-quality translation.
In 2 aspects I might tend to agree with your feeling:
1: The inclusion of "or near-native" (although I might trust a truly native speaker of Indian-English more if she/he is very good at localization, than a near-native speaker of--say--Astralian-English who is not that good at localization).
2: No specification of *which* version of English (UK? US? AU? Kriol? Chinglish?); that shows a level of assumption I don't like.
| | | | | Pages in topic: < [1 2] | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » editing a high quality machine 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 » |
| | Anycount & Translation Office 3000 | Translation Office 3000
Translation Office 3000 is an advanced accounting tool for freelance translators and small agencies. TO3000 easily and seamlessly integrates with the business life of professional freelance translators.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
Translation news
