| Pages in topic: < [1 2 3] | Poll: Do you make use of machine translation options inside your CAT tool(s)? Thread poster: ProZ.com Staff
|
|---|
Lingua 5B 
Bosnia and Herzegovina
Local time: 21:05
Member (2009)
English to Croatian
+ ...
Baran Keki wrote: Ice Scream wrote:
Like using TMs provided by an agency.
Having to use the TM provided by an agency is like wearing someone else's dirty socks, most of the time.
Funny, but true.
In all professions reconstructions and redos cost more than creating something from scratch and most providers avoid the former (eg. construction works, etc.).
Only in translation, let me guess, redos cost less?
| | | | | Two edged sword | Feb 4, 2023 |
Baran Keki wrote:
Speaking of agency TMs/CAT tools, why the hell do agencies have your name displayed in full at the bottom of their CAT tool, revealing your identity to the reviewer (meaning the other translator, who's likely to know who you are through Proz)? Whatever happened to confidentiality, anonymity, 'right to be forgotten' and all that shit? After having signed and scanned all those Confidentiality rubbish for them, they don't extend you the same effing courtesy?
To implement your own name fully in the TM entries by purpose can be a two-edged "sword": On the one hand, you can get the attention of potential clients who one day will have the TM in hand the agency still works with (but now with other, cheaper "colleagues"), and on the other hand, you can fall victim to earlier University classmates of you who have to proofread your work and who will do their best to cut it into pieces, to replace you.
[Bearbeitet am 2023-02-04 11:24 GMT]
| | | | Alex Lichanow
Germany
Local time: 21:05
Member (2020)
English to German
+ ...
Baran Keki wrote:
Having to use the TM provided by an agency is like wearing someone else's dirty socks, most of the time.
My sincerest thanks for this image, it fits at least like a very good sock indeed!
And having your rates reduced based on the matches coming from that TM... Are you supposed to or expected to or do you feel duty-bound to correct/re-translate a 100% match if it's complete and utter rubbish? Or do you simply confirm it and proceed to the next one because you're paid only 10% of your full rate for those matches? And if you do that, would you shift the blame on to the previous/other translator if the client complained of the quality of translation? Or do you take the rap for that?
I will usually correct any glaring misspellings and/or QA errors (those being present at all speaks volumes about the quality consciousness of some "colleagues"), but otherwise skip 100% matches when proofreading my work. Not worth the effort. The best thing about it is that several of my clients know exactly how poor the translations from some of my so-called "colleagues" are, so why they keep working with them at all is anyone's guess.
So, yes, I am actually grateful when 100% matches are not paid to begin with.
[Edited at 2023-02-04 12:36 GMT]
| | | | Baran Keki
Türkiye
Local time: 22:05
Member
English to Turkish
| What I mean by rubbish | Feb 4, 2023 |
Alex Lichanow wrote:
I will usually correct any glaring misspellings and/or QA errors (those being present at all speaks volumes about the quality consciousness of some "colleagues"), but otherwise skip 100% matches when proofreading my work. Not worth the effort. The best thing about it is that several of my clients know exactly how poor the translations from some of my so-called "colleagues" are, so why they keep working with them at all is anyone's guess.
So, yes, I am actually grateful when 100% matches are not paid to begin with.
[Edited at 2023-02-04 12:36 GMT]
What I mean by rubbish is poor style, i.e. 'word-for-word', literal translation that just reads awful, and also the translation mistakes (resulting from that), like for example translating "my takeaway from this discussion" as "my food delivery from this discussion", or "if you have a COPD exacerbation" as "if you own/possess COPD exacerbation" so on and so forth. I come across this kind of stuff almost weekly and they are usually 100% matches, and my clients don't know exactly how poor those matches, because they don't speak/understand a word of Turkish (that's what I've been trying to get across on these forums on many occasions, but I very much doubt the translators of major languages, especially those translating into English, will ever understand this).
So, you skip the 100% matches produced by other translators, even if they're stylistically crap, and your clients are okay with it? That's interesting!
[Edited at 2023-02-05 06:05 GMT]
| | |
|
|
|
Baran Keki wrote:
my clients don't know exactly how poor those matches, because they don't speak/understand a word of Turkish (that's what I've been trying to get across on these forums on many occasions, but I very much doubt the translators of major languages, especially those translating into English, will ever understand this).
Comedy errors are rarely a problem for the end-reader. It’s the subtle ones, the ones that are plausible in the context, that are the real problem. And they are just as undetectable for the agency in translations into English.
As for anyone ignoring 100% matches that are wrong... 😱😱 How incredibly unethical and unprofessional. How hard would it be to tell the customer and get them to pay you to fix them?
| | | | Alex Lichanow
Germany
Local time: 21:05
Member (2020)
English to German
+ ...
| Unethical and unprofessional | Feb 7, 2023 |
Ice Scream wrote:
As for anyone ignoring 100% matches that are wrong... 😱😱 How incredibly unethical and unprofessional. How hard would it be to tell the customer and get them to pay you to fix them?
What I meant to say before becoming annoyed with my job just by spelling out its reality is that I will not correct any poor stylistic choices in 100% matches that are otherwise grammatically and factually correct. Obviously, I will always correct any incorrect facts, such as a missing "not" that would completely twist the source intent.
This might also call for a discussion on how unethical and unprofessional it is to constantly mistranslate one's jobs in the first place. By this, I do not mean the occasional slip-up or "brain fart". What I mean is the constant stream of extremely poor translations that I get to deal with on a near-daily basis in the form of 100% matches.
| | | | Baran Keki
Türkiye
Local time: 22:05
Member
English to Turkish
| Couldn't agree more | Feb 7, 2023 |
Alex Lichanow wrote:
This might also call for a discussion on how unethical and unprofessional it is to constantly mistranslate one's jobs in the first place. By this, I do not mean the occasional slip-up or "brain fart". What I mean is the constant stream of extremely poor translations that I get to deal with on a near-daily basis in the form of 100% matches.
This is exactly what happens when every Tom, Dick and Harry, trusting their good English, tries their hand at translation, armed with the advice they received from Proz forums, with absolutely zero translation experience, skills, knowledge and mastery of their source and target languages.
It seems that the agencies are either sending their rubbish to more seasoned, expert translators to 'revise' or 'proofread' them at the proofreading rates that are closer to the newbie's translation rate (which must be about 3 or 4 cents pw), or hoping that those incompetents will somehow better themselves by learning from Machine Translation (MT training the human translator, if you like, how about that?)
| | | | Samuel Murray
Netherlands
Local time: 21:05
Member (2006)
English to Afrikaans
+ ...
Hans Lenting wrote:
I guess that you are not referring to injecting your preferred terms into MT suggestions?
Google Translate often uses the familiar form of address, but when I need to translate using the polite form of address, four quick find/replace operations on the MT TM fixes it and saves me a bunch of fingerwork later. Also, GT sometimes capitalizes the indefinite article in places where it should be lowercase. But yes, being able to fix terminology errors when I notice it, is useful. I usually don't know beforehand what kinds of mistakes GT is going to make. I suppose if a CAT tool had realtime MT correction functionality, a lot of this could be done using that method instead. My CAT tool, WFC, does have a find/replace feature that performs a series of find/replace operations on TM matches in realtime, but it's not trivial to edit the find/replace list.
Pre-editing the MT TM has other benefits. E.g. I have a client who wants me to pluralize "medicine" only when the source text does it, but GT doesn't always do that, so editing the MT TM by opening it in a richtext editor and using highlighting on the various source and target text alternatives helps me check which segments need to be updated. It does help that my CAT tool, WFC, has a very simple human-readable TM format that you can edit in Word.
And: Is there no way to skip the alignment step?
Err, no. If you want a TM, you have to align. Or, you have to create a source-only TM and then translate the TM, but to do that, you need a program that is capable of doing that. Aligning a set of sentences and their translations is simple and quick (it takes less than a minute).
[Edited at 2023-02-07 09:17 GMT]
| | |
|
|
|
| Re. lists of "creative" suggestions from MT | Feb 8, 2023 |
Would such lists be of any help if they were available within the interface of a CAT tool? I get suggestions from the TB and from Fragment Matches in Trados, and I guess they are about as helpful as this feature can get (despite being 100% mechanical; I wonder if the Muse feature in memoQ does the same thing: I never got around to training the muse). Lists of "creative" translation ideas from MT will still be coming from a machine with no brain. And you'll have to navigate those lists, either by... See more Would such lists be of any help if they were available within the interface of a CAT tool? I get suggestions from the TB and from Fragment Matches in Trados, and I guess they are about as helpful as this feature can get (despite being 100% mechanical; I wonder if the Muse feature in memoQ does the same thing: I never got around to training the muse). Lists of "creative" translation ideas from MT will still be coming from a machine with no brain. And you'll have to navigate those lists, either by reaching for the "Down" key with your pinky or by using your mouse or trackpad to explore what that thing has to offer. Next, you'll have to take your time and decide if there's any value at all in each list item and which one is the best, and then commit the find by pressing "Enter" or whatever. Why not use this time to study the original text and come up with your own ideas, which, at least in my case, will most often be non-linear and non-reproducible by the machine. Often as not, I find myself reaching for the "Escape" key to kill the pesky Autosuggest list (Which Is Often Full of CAPITALIZED ENTRIES to Good Measure, thanks to some people who populated the TM in the past) and return to my normal workflow. All things considered, I guess I wouldn't appreciate any suggestions from MT. I know you can navigate those lists by typing, but they're still hanging there until you commit. What a drag. I downloaded OPUS and tried it on one relatively small document. No surprises there, it's no better (and probably no worse) than Google Translate, I still have to rewrite things and still get no appreciable time gains. Maybe it will become better over time. But again, I'm not hoping it'll ever be able to seamlessly repoduce any non-linear translation decisions I had to make when working on that small project. I have yet to see a piece of machine translation that won't read as completely vacuous content 90% of the time
[Edited at 2023-02-08 09:26 GMT] ▲ Collapse
| | | | finnword1
United States
Local time: 15:05
English to Finnish
+ ...
If I translate a 160,000-word book, yes. If a patent application, (h*ll no).
| | | | Samuel Murray
Netherlands
Local time: 21:05
Member (2006)
English to Afrikaans
+ ...
Samuel Murray wrote:
My CAT tool, WFC, does have a find/replace feature that performs a series of find/replace operations on TM matches in realtime, but it's not trivial to edit the find/replace list.
I don't use this feature, so I misunderstood it.
https://www.wordfast.net/zip/WFC8_manual.html#gloFR
Apparently, it performs the find/replace operation on the text in the active segment when you commit the segment, i.e. not on the TM match. This would be entirely unuseful to me. And in the latest version of WFC, it's relatively trivial to edit the F/R list (although you still need to learn/use a special syntax).
| | | | Tom in London
United Kingdom
Local time: 20:05
Member (2008)
Italian to English
| TMs provided by agencies | Feb 10, 2023 |
I was recently sent a TM that the agency asked me to use. But the terminology was all wrong (to be clear: *w r o n g* - not different from the terminology I would have used).
I told the agency and they insisted that I use the TM they had sent me.
I withdrew from the job.
If I'm translating something, it has to be MY translation. Otherwise I simply won't do it.
[Edited at 2023-02-10 10:38 GMT]
| | | | | Pages in topic: < [1 2 3] | 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 make use of machine translation options inside your CAT tool(s)? | 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 » |
| | 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 
