Pages in topic:   [1 2] >
Translation theft ??
Thread poster: walkabout
walkabout
walkabout  Identity Verified
Spain
Local time: 16:11
English to Spanish
+ ...
Oct 31, 2014

After the revelations by whistleblowers such as Assange and Snowden, it has become apparent that there are some big internet companies out there which have extremely easy access to our data. In fact, they are the owners of almost all the networks and servers which our data either crosses through or resides on.

Some of these companies, search engines, etc are very familiar to us all (no need to mention any names). A good number of them have automated translation software, which at th
... See more
After the revelations by whistleblowers such as Assange and Snowden, it has become apparent that there are some big internet companies out there which have extremely easy access to our data. In fact, they are the owners of almost all the networks and servers which our data either crosses through or resides on.

Some of these companies, search engines, etc are very familiar to us all (no need to mention any names). A good number of them have automated translation software, which at the moment doesn't pose a real threat to a human translator given its poor results... for now. However, for those who have followed this software's development over the years, it can be said that it has improved in leaps and bounds... and it continues to do so. Obviously there are skilled programmers behind these automated translators making them a little better each day with hard work and good programming, but is there another way to supplement their efforts?

In light of the above-mentioned revelations, and after doing a little research, I've come to suspect that our (I'm referring to us human translators) hard work, the fruit of endless evenings and nights at the keyboard, might be being intercepted when it is emailed. To be more exact, our bilingual files (.sdlxliff) are probably being automatically intercepted when we email them back to our clients. Once intercepted they can be easily incorporated into any automated translation software database.

Maybe I'm a little paranoid... or maybe not; perhaps no one is intercepting these bilingual files and making dishonest use of them, or perhaps they are. Whatever the case may be, I prefer to send my finished product with some type of security, just as when I post a letter, I like to seal the envelope - it would never occur to me to post it unsealed.

I send my completed bilingual files back to my clients/agencies inside an encrypted ZIP file that is only openable with a pre-shared password. After a certain time (maybe a month or two) we change the password.

I wouldn't dream of telling anyone how to go about their business, but I'd recommend that we all considered this measure.

[Edited at 2014-10-31 18:17 GMT]
Collapse


 
Robert Rietvelt
Robert Rietvelt  Identity Verified
Local time: 16:11
Member (2006)
Spanish to Dutch
+ ...
Intercepting? Oct 31, 2014

No, I don't think so, but I am sure that they are making use of our translation skills somehow. As you said: "it has improved in leaps and bounds... and it continues to do so.", I noticed that too.

What happens with the TM's you deliver together with your translations? Are we digging our own graves? For survey's and other "official" forms, certainly. I almost never receive a job in this field anymore, and when I do, it is finished within a few clicks, I got it all in my TM's.
... See more
No, I don't think so, but I am sure that they are making use of our translation skills somehow. As you said: "it has improved in leaps and bounds... and it continues to do so.", I noticed that too.

What happens with the TM's you deliver together with your translations? Are we digging our own graves? For survey's and other "official" forms, certainly. I almost never receive a job in this field anymore, and when I do, it is finished within a few clicks, I got it all in my TM's.

For translations, where style and actuality count, never, that is to say, not in my lifetime. The first computer that "feels" still has to be invented. Starship Enterprise is still exploring other galaxies.

[Edited at 2014-10-31 17:46 GMT]

[Edited at 2014-10-31 18:00 GMT]
Collapse


 
Thayenga
Thayenga  Identity Verified
Germany
Local time: 16:11
Member (2009)
English to German
+ ...
Not in our lifetime Nov 1, 2014

As Robert has stated, these files are unlikely to be intercepted by anyone. And there are still several fields, e. g. journalism, multi-media, literature, in which automated translation will never work, that is not in our lifetime.

However, as I had mentioned in another thread, the delivered TM's are being put to use, to save the agency expensive human translations in certain fields. They simply use what had been previously translated.

It's always possible that parts of
... See more
As Robert has stated, these files are unlikely to be intercepted by anyone. And there are still several fields, e. g. journalism, multi-media, literature, in which automated translation will never work, that is not in our lifetime.

However, as I had mentioned in another thread, the delivered TM's are being put to use, to save the agency expensive human translations in certain fields. They simply use what had been previously translated.

It's always possible that parts of these translations end up on those no-need-to-name-them sites.
Collapse


 
walkabout
walkabout  Identity Verified
Spain
Local time: 16:11
English to Spanish
+ ...
TOPIC STARTER
Translation theft ?? Nov 1, 2014

How can you assert that this is "unlikely"? We all thought it was "unlikely" the USA was spying, monitoring and storing every one of our communications, be they written or verbal, and they are. We all thought it was "unlikely" that the biggest financial crisis of our times was about to occur, and it did. We all think it is "unlikely" that our daughter could be the slut of the school, etc... Human nature - I guess we all need to believe everything is fine.

 
Phil Hand
Phil Hand  Identity Verified
China
Local time: 23:11
Chinese to English
Definitely accessed, not a worry Nov 1, 2014

Walkabout is right that the files we email back and forth are probably accessed by Google and other large companies. Remember the selling point of Gmail is that it *reads your emails and gives you relevant adverts*. To imagine that attachments are off limits is just naive.

But I find it hard to worry about this because most of my product ends up accessible to Google anyway. Quite a lot of my translation output is for publication in books or journals, and the source texts are mostly
... See more
Walkabout is right that the files we email back and forth are probably accessed by Google and other large companies. Remember the selling point of Gmail is that it *reads your emails and gives you relevant adverts*. To imagine that attachments are off limits is just naive.

But I find it hard to worry about this because most of my product ends up accessible to Google anyway. Quite a lot of my translation output is for publication in books or journals, and the source texts are mostly available online as well. That means that Google already has the option of creating TMs out of my work without bothering to snoop on my email.

So I'm not too fussed. As the various machine translation services show, no volume of word juggling and calculomatix will ever constitute real translation, for a very simple reason: to translate a text properly, you have to know what it means. Now, one day computers will start to be able to read texts, and on that day we're out of a job - but so is half the world.
Collapse


 
Jeff Whittaker
Jeff Whittaker  Identity Verified
United States
Local time: 11:11
Member (2002)
Spanish to English
+ ...
The myth of translation memory Nov 1, 2014

I think that even if we had storage capacity large enough to hold every single bilingual text/translation memory on the planet and computing power fast enough to quickly process all of this data, we would find that the number of "matches" would not be significantly higher beyond what is currently achieved at the inter-company level (translating documents created by the same company). The nature of human communication and language is such that the possible combinations are limitless and we are al... See more
I think that even if we had storage capacity large enough to hold every single bilingual text/translation memory on the planet and computing power fast enough to quickly process all of this data, we would find that the number of "matches" would not be significantly higher beyond what is currently achieved at the inter-company level (translating documents created by the same company). The nature of human communication and language is such that the possible combinations are limitless and we are always creating new ideas.

That's one of the reasons that I find "collecting" translation memories a complete waste of time.

It's analogous to those people who spend inordinate amounts of their time clipping and collecting coupons in order to go to the grocery store to purchase 500 boxes of breakfast cereal, 100 jars of spaghetti sauce, 50 bottles of ketchup (only in America?). They are proud of the fact that they "saved" $1,000 or got $1,000 worth of groceries for free. However, once they deduct from this "savings" the amount of time they spent collecting the coupons, scouring the advertisements for sales, purchasing the items, and deduct all of those grocery items that will never ever use or that will spoil because they will never get a chance to use them, their actual savings is close to zero.

It's the same with CAT tools. Unless you do a lot of work for the same company, the amount of time you save by not having to re-translate a odd page or paragraph must be offset by the cost of the tool itself, the time spent using the tool, dealing with the errors and bugs and managing the memories, the extra storage capacity, the fact that you may not be paid for these items (due to "discount schemes"), etc. must be compared to the time it would take to just retranslate the text or offset by your ability to say "yes, I remember translating that before" and doing a simple search through your documents (if you save them) AND GET PAID FOR IT. Actual time savings and monetary gain = zero.

And all of this does not take into consideration the fact that all "memories" will not be perfect as some will undoubtedly contain errors and mistakes. So, let them collect all the "memories" they want - it won't make a bit of difference.

What may disappear in the future (and God I hope so) will be the need to translate what will become standardized digitized personal documents (birth certificates, school transcripts, etc.).

[Edited at 2014-11-01 14:43 GMT]
Collapse


 
Thayenga
Thayenga  Identity Verified
Germany
Local time: 16:11
Member (2009)
English to German
+ ...
It depends Nov 1, 2014

walkabout wrote:

How can you assert that this is "unlikely"? We all thought it was "unlikely" the USA was spying, monitoring and storing every one of our communications, be they written or verbal, and they are. We all thought it was "unlikely" that the biggest financial crisis of our times was about to occur, and it did. We all think it is "unlikely" that our daughter could be the slut of the school, etc... Human nature - I guess we all need to believe everything is fine.


Well, it's obvious (or should be) that all governments (not just the USA) are observing others, including private individuals. It's been this way throughout history - even without the www -, meaning I never considered that "observation" to be unlikely knowing about it being/having been in practice all the while.

The financial crisis was also predictable. Has been for years before it finally occurred. And no, it wasn't "unlikely".

That Google is reading all correspondence and monotoring all activities on the Web was known even before they finally had to admit it. This can be "somewhat" avoided by not using neither Gmail nor any other service (?) provided by Google.

walkabout wrote:
I guess we all need to believe everything is fine.


IMO there is quite a difference between "believing everything's fine" and being realistic and seeing the risks as they are without adding "extra fuel" or possibly "blowing things out of proportion".

All these are serious issues that can, under certain circumstances, lead to paranoia. And please note that I am not even suggesting that you might be, okay? Not only our profession but life itself are full of risks, taking chances every day. So of course nothing is ever "unlikely".


 
Daryo
Daryo
United Kingdom
Local time: 15:11
Serbian to English
+ ...
Translation theft ? no real need for that Nov 1, 2014

"I send my completed bilingual files back to my clients/agencies inside an encrypted ZIP file that is only openable with a pre-shared password. After a certain time (maybe a month or two) we change the password."

Sorry to shatter your illusions, but the likes of NSA and GCHQ (and not only them) have at their disposal the kind of computing power that will break your ZIP password in a blink. Not to mention all other potential weak links between you and the final client...

... See more
"I send my completed bilingual files back to my clients/agencies inside an encrypted ZIP file that is only openable with a pre-shared password. After a certain time (maybe a month or two) we change the password."

Sorry to shatter your illusions, but the likes of NSA and GCHQ (and not only them) have at their disposal the kind of computing power that will break your ZIP password in a blink. Not to mention all other potential weak links between you and the final client...

"I wouldn't dream of telling anyone how to go about their business, but I'd recommend that we all considered this measure"

now, that start sounding slightly paranoid, and actually not much useful -more likely to give an impression/illusion of privacy than actually achieving it.

BTW, you seem to have forgotten terabytes (if not petabytes or exabytes) of publicly available multilingual documents and various online glossaries, dictionaries encyclopedia etc. ; anyone with enough bandwidth and processing power can collect and analyse them. THAT seems to me a far more likely source to be used to improve MT output...
Collapse


 
José Henrique Lamensdorf
José Henrique Lamensdorf  Identity Verified
Brazil
Local time: 12:11
English to Portuguese
+ ...
In memoriam
Accountability & liability Nov 1, 2014

I see hundreds, maybe thousands of translators using free e-mail services, like Gmail, Hotmail, etc. AFAIK none of these offers any kind of formal guarantee that the contents will not be disclosed, nor do they charge anything from users in exchange for their services.

Gmail is under the same umbrella as Google Translate. Hotmail is under the same as Bing/MS Translator.

It would be easy to spot and focus on translators who do it by matching their web sites, online portal
... See more
I see hundreds, maybe thousands of translators using free e-mail services, like Gmail, Hotmail, etc. AFAIK none of these offers any kind of formal guarantee that the contents will not be disclosed, nor do they charge anything from users in exchange for their services.

Gmail is under the same umbrella as Google Translate. Hotmail is under the same as Bing/MS Translator.

It would be easy to spot and focus on translators who do it by matching their web sites, online portals, etc. with their e-mail addresses.

What would prevent Google/MS to develop some contrivance that would parse files going in/out such translator e-mail addresses on their servers, and use them to feed their machine translation databases? Apparently Linguee and others do it with bi-/multi-lingual web sites, so the technical solution is there already.

I have been using a paid e-mail service since day one. It's Mandic, in Brazil, and there should be thousands just like it everywhere. Though I know personally the founder, Aleksandar Mandic, he has sold it - lock, stock, and barrel - to an investment bank. I pay something like USD 5.00 per month, however we have a contract, whereby they are liable for non-disclosure. If anything leaks from my e-mail, they will be accountable and liable for megabucks; above all, they know it!

This article should be enlightening:
http://edition.cnn.com/2014/03/31/tech/web/gmail-privacy-problems/
Collapse


 
walkabout
walkabout  Identity Verified
Spain
Local time: 16:11
English to Spanish
+ ...
TOPIC STARTER
Translation theft ?? Nov 1, 2014

I agree with you José, "when the product is free, it's because YOU are the product"

 
Merab Dekano
Merab Dekano  Identity Verified
Spain
Member (2014)
English to Spanish
+ ...
Don't worry, be happy... Nov 1, 2014

Almost every day I do small jobs, such as translations of articles, press releases, speeches, etc. It has been a while I am doing it and I am yet to see any tangible benefits (if any) from my TM, except, of course, dates, numbers and similar.

Creative texts are out there to be translated by humans, no matter how "sophisticated" MTs get.

Protecting our output is unreasonable. I am not saying it is improper, but it is, for sure, unrealistic. So, why bother about something
... See more
Almost every day I do small jobs, such as translations of articles, press releases, speeches, etc. It has been a while I am doing it and I am yet to see any tangible benefits (if any) from my TM, except, of course, dates, numbers and similar.

Creative texts are out there to be translated by humans, no matter how "sophisticated" MTs get.

Protecting our output is unreasonable. I am not saying it is improper, but it is, for sure, unrealistic. So, why bother about something that is already way beyond our control?
Collapse


 
jyuan_us
jyuan_us  Identity Verified
United States
Local time: 11:11
Member (2005)
English to Chinese
+ ...
it is not necessary to do so. Nov 1, 2014

walkabout wrote:

I send my completed bilingual files back to my clients/agencies inside an encrypted ZIP file that is only openable with a pre-shared password. After a certain time (maybe a month or two) we change the password.


Don't you think this practice is an extra burden for both your clients and yourself?

I think we needn't be so sensitive to something that is beyond our control, or for which we needn't worry about.

As a business owner, you translate, and you get paid for what you have translated. That's all.


 
José Henrique Lamensdorf
José Henrique Lamensdorf  Identity Verified
Brazil
Local time: 12:11
English to Portuguese
+ ...
In memoriam
Very smart Nov 1, 2014

jyuan_us wrote:

I think we needn't be so sensitive to something that is beyond our control, or for which we needn't worry about.

As a business owner, you translate, and you get paid for what you have translated. That's all.


Some clients have translators sign a very strict NDA, wait until snail-mail delivers it to them in hard copy, and then they send out the job... on Google Docs, without any password, just the link to download.


 
nrichy (X)
nrichy (X)
France
Local time: 16:11
French to Dutch
+ ...
Noticed that too Nov 1, 2014

José Henrique Lamensdorf wrote:

Some clients have translators sign a very strict NDA, wait until snail-mail delivers it to them in hard copy, and then they send out the job... on Google Docs, without any password, just the link to download.


or by Wetransfer and the like. The other day I downloaded files from the agency's Dropbox: they were there amidst other files.

If a client sends work to a translator by e-mail, then the translator can send it back in the same way. And if the client sends something to his translator knowing that he or she uses gmail or hotmail, then he takes the risk that this free service is more sensitive to hacking than another one.

Although I am one of those who does not use cloud based services for this reason, I also think that there's no need to be overly paranoid. If the client takes a risk, we can take the same risk. And: files that are really secret are translated inhouse in the company, and in some cases (army etc.) on a computer that is not connected to the internet.


 
Robert Rietvelt
Robert Rietvelt  Identity Verified
Local time: 16:11
Member (2006)
Spanish to Dutch
+ ...
Aren't we exaggarating? Nov 3, 2014

Are we loosing it, or is this normal behaviour now? Do we see goshts everywhere or what?

I send my TM's to my clients (if they wish so). Maybe someone ('THEM') is intercercepting it, maybe not. Both my Dutch translation as well as the orginal version are available on the web, and with a little 'alligning', all is avaible to anyone.

So what are we talking about?



[Edited at 2014-11-03 23:54 GMT]


 
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 »

Translation theft ??







Trados Studio 2022 Freelance
The leading translation software used by over 270,000 translators.

Designed with your feedback in mind, Trados Studio 2022 delivers an unrivalled, powerful desktop and cloud solution, empowering you to work in the most efficient and cost-effective way.

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 »