| Pages in topic: < [1 2 3 4 5 6 7 8 9 10] > | Clients / large translation companies now talking about pooling linguistic data. Should we be there? Thread poster: Henry Dotterer
|
|---|
| TAUS, reliable or not? | Jul 31, 2008 |
Dear All,
I have read all the posts (answers and/or questions) for this interesting issue and I feel this question has been somewhat overlooked: would TAU project be realiable and/or useful?
The answer to this question is nor a simple yes or not, since we are speaking of a very large project of pooling linguistic data from different TMs.
Some people have mentioned in their posts their fear that such a big project would lack quality and I have to agree with... See more Dear All,
I have read all the posts (answers and/or questions) for this interesting issue and I feel this question has been somewhat overlooked: would TAU project be realiable and/or useful?
The answer to this question is nor a simple yes or not, since we are speaking of a very large project of pooling linguistic data from different TMs.
Some people have mentioned in their posts their fear that such a big project would lack quality and I have to agree with them. TAU will probably create some "standard quality level" for their final product, as it would be stupid not to. But, who will decide this standard? TAU big boys? Translation associations?
Language in general is very ambiguous, even specialized languages such legal, technical or scientific languages. Who decides which translation is the correct one? It is very difficult to take such a decision. And in any case, it will be only useful for a given number of cases, as it will not take into account the different contexts in which you can find a similar "expression/sentence".
In order to create a reliable big data pool for automatic translation, every one should agree what is the minimum translation unit, a sentence, fixed expressions or even a whole text can have very different interpretations, and therefore translations, depending upon their context, the tone of the document, the audience the text is directed to, etc.
Linguists, translators, translation agencies, translation associations and all professionals related with the translation industry would need to reach an agreement in terms of this minimum translation unit and quality assurance first before beginning such a large project.
I also think, some might disagree with my opinion, that such a big project could be confusing for its users unless they are able to create a very good index and/or search engine.
Because it will use so many different TMs, some of them for different language pairs, different specializations. I would need to create a clear categorization (another complex issue) for its index and a good search engine which should be fine tuned so its querys are as relevant as possible. Meaning, me as an user create a query and the search engine should return a relevant answer to my query otherwise, its answers would be useless or even confusing.
For this project to work, it would need the participation of all sorts of experts, not only translators, linguists to ensure that it has a good quality. But also good coordinators and good moderators, so there is a good comunication among the diffferent groups that form part of the project. It would need IT and computer science experts so the final tool is easy to use, and the results it has are relevant to its final user. It would probably need editors, copywriters and legal to sift all copywrited, sensible data from the TMs and leave a good, understandable data at the end. There is probably a bunch of other professionals and specialist which would be needed for such a project.
I personally don't see this happening any time soon, since the translation industry, as some have already stated, is too fragmented. There are not agreements event in the most essential issues, such as quality. Therefore, this project would have a very limited use, only for repetitive texts or versions of a previous document.
I don't think it would it change the translation industry as it is, because many people would find just not usefull for their business. Even some Translation Agencies, who pride themselves on their quality, would not use such a tool because they wouldn't trust it.
In any case, I believe it to be a good idea that could, in the future, help translators to provide a better quality and/or unified final product for their clients. But it will never replace the human translator who will have to decide whether one of its results is useful for his/her given project. ▲ Collapse
| | | | Tomi Dobaj 
Slovenia
Local time: 17:28
Member (2007)
English to Slovenian
+ ...
SITE LOCALIZER | Under the flag of Henry... | Jul 31, 2008 |
In any case, whether or not the corporations' intentions are evil, it would be wise to know as much as possible about their "pooling" plans.
As they will carry them out in any case, ProZ.com should definitely stay close "at the bowl", so we keep ourselves well informed about the developments in our industry, and maybe even make them share some corporate breadcrumbs with us:)
The idea is really interesting, but the implementation could be really complicated.. I already h... See more In any case, whether or not the corporations' intentions are evil, it would be wise to know as much as possible about their "pooling" plans.
As they will carry them out in any case, ProZ.com should definitely stay close "at the bowl", so we keep ourselves well informed about the developments in our industry, and maybe even make them share some corporate breadcrumbs with us:)
The idea is really interesting, but the implementation could be really complicated.. I already have lots of questions about it.
How would it work in practice, if ProZ.com joins up?
Would be members, who decide to share their TMs, given an access to the whole/a part of the "pool"? Would some members really "flee" because of that?
How about supervision of the shared, sharers and access? Could the pool be made open-source, or should it be closed to the proportion one shares in it?
How would it be settled legally (in general)? Are any patent claims possible for such pools?
I wouldn't mind at all if Henry goes there "on our behalf" and see how the pool is being dug and whether we should jump into it. Who else would?
Cheers all.
Tomi ▲ Collapse
| | | | | This won't work | Jul 31, 2008 |
From my experience in working as a sub-subcontractor for the *big boys'* regional translation centres, they have a few things in common:
1. Their standpoint is that their own termbase is a holy writ that cannot be criticised by any blasphemous outsider;
2. The terminology they use is the only right one in the whole wide world and if any competitor uses a different one - well, his problem...
3. Anything that is not strictly technical but included in their TM is of invariably low... See more From my experience in working as a sub-subcontractor for the *big boys'* regional translation centres, they have a few things in common:
1. Their standpoint is that their own termbase is a holy writ that cannot be criticised by any blasphemous outsider;
2. The terminology they use is the only right one in the whole wide world and if any competitor uses a different one - well, his problem...
3. Anything that is not strictly technical but included in their TM is of invariably low quality or produces a low-quality effect in the final text. For instance, there is only a single acceptable equivalent of *therefore* in any given language - they want a machine-like translation, so they get it. Additionally, #1 and #2 apply to those items as well.
I can't see how this widespread approach might lead to any compromise, alignment etc.
Therefore I can't see any reason for getting excited about the concept - we'll soon see its agony in the midst of its unyielding parents. Even if they could earn a few pennies more by keeping it alive and kicking. They could possibly overcome the technical / organisational problems with their vast resources. The way of thinking is much harder to remodel.
[Edited at 2008-07-31 11:24] ▲ Collapse
| | | | Anthony Baldwin
United States
Local time: 11:28
Portuguese to English
+ ...
| A translation is a translation, and if it is supposed to be good work, it needs a human translator. | Jul 31, 2008 |
AFLSSInc wrote:
I couldn't agree more. There is no doubt in my mind that the job description of "Translator" will change a bit as technology progresses. In the end, this industry is just starting to catch up on technology trends other industries went through 15 years ago.
A translation is a translation, and if it is supposed to be good work, it needs a human translator. Further, as mentioned before, I don't think fully automated translation won't be possible for another few generations of translators.
Regards
Paul
I don't believe the human factor can or will ever be removed from the translation equation, since human language is dynamic, ever changing, and, even with the internet bringing us ever closer, cultural differences persist which only a human can comprehend fully, never a computer.
That said, clearly technology is streamlining our work, and new technologies are both, creating more work for us in some areas, reducing our potential work load in others.
What it means is, to survive in this industry, we need to be on top of the technological advances.
A typewriter and a Collins-Robert haven't been sufficient for some years.
Already a word processor and en e-mail account aren't either.
I've noticed certain trends, such as, companies for whom I've worked for some years are now sending me less translations, but more revision work, where their translation quite often was clearly generated from work I, myself, had previously done, no doubt with translation memory technology. It's still work.
And there are new clients continuously appearing with needs that are not met by large translation agencies with huge banks of memories, either, due to the content of their material, or, sometimes simply the formats in which their product is available and/or needed.
We need to be aware of where new markets are emerging, and adjust our strategies to meet their needs. We need to be aware of technological advances and make them a part of what we do.
So long as we do these things, we will still be needed.
| | |
|
|
|
Umberto Steindler (X)
Local time: 17:28
German to Italian
+ ...
| Social and juridical responsibility | Jul 31, 2008 |
Dear Henry
The professional issues related to our profession may be very important, but there are other issues, which may affect the entire society.
1. Are you sure that no legal objection to the initiative is possible?
2. Are you sure that no intelligence agency or terror organisation can infiltrate a person or penetrate with hackers in this new institution and use reverse engineering to recover the complete documentation on confidential projects?
Yours sincerely
... See more Dear Henry
The professional issues related to our profession may be very important, but there are other issues, which may affect the entire society.
1. Are you sure that no legal objection to the initiative is possible?
2. Are you sure that no intelligence agency or terror organisation can infiltrate a person or penetrate with hackers in this new institution and use reverse engineering to recover the complete documentation on confidential projects?
Yours sincerely
Alberto Valle ▲ Collapse
| | | | Tom in London
United Kingdom
Local time: 16:28
Member (2008)
Italian to English
| maybe i'm cynical | Jul 31, 2008 |
it seems to me this is simply a way for a lot of powerful companies to force down their translation costs by creating a cartel or monopoly. The losers in this process would be the translators.
| | | |
I mean I have been inthe market for a while and for what I can see translation goes beyond the limits of passing from one language to another, I mean it needs interpretation as well as structure. This may help but a good, reliable human transaltor will always be needed.
AFLSSInc wrote:
I couldn't agree more. There is no doubt in my mind that the job description of "Translator" will change a bit as technology progresses. In the end, this industry is just starting to catch up on technology trends other industries went through 15 years ago.
A translation is a translation, and if it is supposed to be good work, it needs a human translator. Further, as mentioned before, I don't think fully automated translation won't be possible for another few generations of translators.
Regards
Paul
I don't believe the human factor can or will ever be removed from the translation equation, since human language is dynamic, ever changing, and, even with the internet bringing us ever closer, cultural differences persist which only a human can comprehend fully, never a computer.
That said, clearly technology is streamlining our work, and new technologies are both, creating more work for us in some areas, reducing our potential work load in others.
What it means is, to survive in this industry, we need to be on top of the technological advances.
A typewriter and a Collins-Robert haven't been sufficient for some years.
Already a word processor and en e-mail account aren't either.
I've noticed certain trends, such as, companies for whom I've worked for some years are now sending me less translations, but more revision work, where their translation quite often was clearly generated from work I, myself, had previously done, no doubt with translation memory technology. It's still work.
And there are new clients continuously appearing with needs that are not met by large translation agencies with huge banks of memories, either, due to the content of their material, or, sometimes simply the formats in which their product is available and/or needed.
We need to be aware of where new markets are emerging, and adjust our strategies to meet their needs. We need to be aware of technological advances and make them a part of what we do.
So long as we do these things, we will still be needed. [/quote]
| | | |
Tom in London wrote:
it seems to me this is simply a way for a lot of powerful companies to force down their translation costs by creating a cartel or monopoly. The losers in this process would be the translators.
But if we continue accomplishing the tasks with quality we will always have our space assured in the working market, don't you think?
| | |
|
|
|
Cath St Clair (X)
Spain
Local time: 16:28
French to English
+ ...
| Increased automation = Mindless work | Jul 31, 2008 |
For me, the problem with increased automation is the mindlessness of editing it.
Having translated for a major automotive company for several years, I recently applied to an agency who were translating for the same company, but using "new, innovative methods that are the future of the translations business" or something such like. The translations were automated, and my job was basically to proof and correct any mistakes.
I accepted my first job, and within a day was bo... See more For me, the problem with increased automation is the mindlessness of editing it.
Having translated for a major automotive company for several years, I recently applied to an agency who were translating for the same company, but using "new, innovative methods that are the future of the translations business" or something such like. The translations were automated, and my job was basically to proof and correct any mistakes.
I accepted my first job, and within a day was bored to tears. It took me a week to complete, and on being offered another, had no hesitation in replying that no, this kind of work was not for me. Admittedly, it was a car manual, but I have translated plenty of manuals before, without pulling my hair out.
I translate because I find it stimulating and challenging to express in my own language what I read in another. That experience helped me decide that my plan for the future would be to move away from the technical arena (in which there is never a shortage of work for my language pairs) and more towards areas that I genuinely enjoy translating (and which are less susceptible to becoming automated). The problem with this, is that there is less demand, and plenty of translators who want to do the same. I have had to strike a balance.
The experience gave me a chill up my spine. If this really is the future of translating, then I'll be looking around for another profession - at least a part time one, so as to ease the monotony. ▲ Collapse
| | | | | I see it like this: | Jul 31, 2008 |
This assumption:
AFLSSInc wrote:
a) Others deliver bad translations that subsequently I have to deal with in some form
If there were a centralized, cross-industry TM repository, openly accessible, backed by big-corp funding, consideration will certainly be given to the QA aspect.
is countered by:
Iza Szczypka wrote:
From my experience in working as a sub-subcontractor for the *big boys'* regional translation centres, they have a few things in common:
1. Their standpoint is that their own termbase is a holy writ that cannot be criticised by any blasphemous outsider;
2. The terminology they use is the only right one in the whole wide world and if any competitor uses a different one - well, his problem...
- which I agree with - and:
Telperie wrote:
But, who will decide this standard? TAU big boys? Translation associations? ... In order to create a reliable big data pool for automatic translation, every one should agree what is the minimum translation unit, a sentence, fixed expressions or even a whole text can have very different interpretations, and therefore translations, depending upon their context, the tone of the document, the audience the text is directed to, etc. ... There are not agreements event in the most essential issues, such as quality.
And again, the common but incisive points regarding matches:
cameronaj wrote:
Has anyone ever come across the perfect TM? Has anyone ever encountered a consistent TM? Those who think that a word in one language should always be translated by the same word in the other language clearly know nothing about language.
and:
Salvador Scofano and Gry Midttun wrote:
100% Matches - TM datababases
Has someone already received TMs with 100% matches and had fun with the brilliant solutions received as "100% matches"? Then imagine this in a database made up with several TMs from several translators. Maybe some were good in the specific contexts, but disastrous in different situations.
Which lead to this:
Cath Murray wrote:
...using "new, innovative methods that are the future of the translations business" or something such like. The translations were automated, and my job was basically to proof and correct any mistakes.
The issue with having some kind of ginormous searchable TM with matches is not only whether or not this can be used as a resource - yes or no - it is HOW it will be extended for use and under what conditions and requirements.
The problem that many of us are facing is not simply that we are correcting others' work, bad quality translations, bad TMs, etc. - and we do face this problem - it is the fact that we are being requested to do so for less money because the source says this is a "match" or a "repetition" and therefore unilaterally decides to lower the value of the work.
Based on what I read (above), I think many would agree with me - the work we do is NOT less valuable because someone else did a different version of it in a different circumstance in a different context in a different language variant for a different customer at a different time.
The question that I feel is particularly germane is this: If this project becomes reality, what can we as translators do to defend this position?
As for the original question of whether Proz.com should become involved, I do not have an answer, because I still am not sure I understand the motive for doing so, the potential (or eventual) outcome of that involvement, and even if I did, my trust is rather grudgingly granted these days.
See above.
[Edited at 2008-07-31 15:18]
| | | | Abba Storgen (X)
United States
Local time: 10:28
Greek to English
+ ...
| In plain English | Jul 31, 2008 |
"...sharing parallel language data with the objective to stimulate innovation and automation of translation activities..."
That sounds like corporate talk for 19 year old secretaries. In plain English:
Large agencies would like to have larger translation memories to cut down on the cost of the translation. That's all. The industry is heading to the elimination of most translators (some will still be around for a few years, to edit and oversee quality), and to the elimin... See more "...sharing parallel language data with the objective to stimulate innovation and automation of translation activities..."
That sounds like corporate talk for 19 year old secretaries. In plain English:
Large agencies would like to have larger translation memories to cut down on the cost of the translation. That's all. The industry is heading to the elimination of most translators (some will still be around for a few years, to edit and oversee quality), and to the elimination of small agencies.
It's a typical phenomenon and it happened in the accounting industry, the real estate industry, the insurance industry, the publications industry, and every other industry that went online. It's happening fast in this industry as well. I give it another 2-3 years before prices are down by another 40%, and more than 70% of translators are gone ("positive spirit" and all the other post-hippy era messages don't apply when your account is empty). I would also dare to predict the elimination of 80-90% of agencies (most of the small ones).
It started with the prices (I predicted the fall 5 years ago, when everything looked fine) and it will spread to vast translation memories used extensively by large agencies.
Statements such as "you need the human factor for quality" remind me of the statements of all other professionals in all other industries who became obsolete by machines.
All in all, that's not so bad. We'll find something else to do... everyone did. ▲ Collapse
| | | | |
|
|
|
Mohamed
United States
Local time: 09:28
English to Arabic
+ ...
| What about my assets? | Jul 31, 2008 |
I really have a question and I hope anyone here in this forum can answer it.
Are we going to benefit from sharing our translation memories and terminology lists? As a translator, translation memories and terminology are my main assets and I should get benefit from sharing them, unless I am doing this voluntarily to help the translators' community.
Isn't copy-righted? Shouldn't we get paid for that? I think we shou... See more I really have a question and I hope anyone here in this forum can answer it.
Are we going to benefit from sharing our translation memories and terminology lists? As a translator, translation memories and terminology are my main assets and I should get benefit from sharing them, unless I am doing this voluntarily to help the translators' community.
Isn't copy-righted? Shouldn't we get paid for that? I think we should - as a community - think about this and see how we can protect our assets. It is really unfair that big companies such as SDL or LionBridge would give my translations and work to anyone else (except for their end client), just for their own benefit.
As a result, we will end up working as editors and we will mainly edit machine-translated text. For sure we will get less and we won't be able to survive as translators, especially with the weak US dollar and sky-rocketing prices for food, gas, real estate, education, etc.
I really do not like TAUS idea of automoation because it really does not even consider the translators' community and we won't get any benefit out of it. Maybe I am wrong and it is not clearly defined yet, but I believe this will have a significant impact on our profession. Sharing our assets with TAUS should be illegalized, I believe. We should as a community, either talk to our clients or at least get some benefits out of this 'initiative'.
[Edited at 2008-07-31 17:08]
[Edited at 2008-07-31 19:42] ▲ Collapse
| | | | Abba Storgen (X)
United States
Local time: 10:28
Greek to English
+ ...
You' re not talking to lawyers or financial advisors here, or accountants, or anyone else who likes to make good money. You' re just talking to a bunch of young folks (or folks with wealthy and easy going spouses), who happen to know a foreign language and they get paid to type stuff. Or at least, that's how they price themselves in the market. "Assets"? Most of your collegaues will work for 10 bucks per hour to help the large agencies build vast translation memories just by alinging text with T... See more You' re not talking to lawyers or financial advisors here, or accountants, or anyone else who likes to make good money. You' re just talking to a bunch of young folks (or folks with wealthy and easy going spouses), who happen to know a foreign language and they get paid to type stuff. Or at least, that's how they price themselves in the market. "Assets"? Most of your collegaues will work for 10 bucks per hour to help the large agencies build vast translation memories just by alinging text with Trados. That's because they would like to keep themselves and their countries poor. They' re used to that.
Do you seriously think that your little databases are "assets" when compared to the allignment of 5-10 million words per month in vast translation memories? ▲ Collapse
| | | | | Get involved. Create a new pricing model | Jul 31, 2008 |
As always, many of us are talking about the translation industry as if it were a single entity rather than a hugely fragmented collection of segments and subsegments scattered across the globe.
Looking at the TAUS member list, and on the basis of my own experience, I'd say that the initiative is designed by and for the big global software and IT houses and is unlikely to have much impact outside that field.
Software localisation is already a long way down the path towar... See more As always, many of us are talking about the translation industry as if it were a single entity rather than a hugely fragmented collection of segments and subsegments scattered across the globe.
Looking at the TAUS member list, and on the basis of my own experience, I'd say that the initiative is designed by and for the big global software and IT houses and is unlikely to have much impact outside that field.
Software localisation is already a long way down the path towards automation and is only likely to go further. Admittedly it is a huge market, but it is certainly not representative of the translation market as a whole.
It's true that those involved in software localisation will have to be prepared to do more editing than translating, but that's an option they're free to accept or refuse.
Anyway, the real point that needs to be addressed here is the pricing model. The key piece of data is always earnings per hour, not per word. It doesn't matter if a translator/editor produces 500 delivery-quality words per hour or 5000; what matters is that they earn an amount for that hour that they feel reflects the value they contribute to the end product.
Judging from comments on Proz.com and elsewhere, I think we can safely say that a significant proportion of translators feel they're getting a raw deal under the current match-based pricing model.
So, in terms of suggestions, I'd propose that Proz.com should:
a) Get involved in the TAUS Data Association to represent its members.
b) Start working with members to develop a new automation-based pricing model that translators, agencies and end clients are all happy with.
The potential exists to create a win-win situation in which translators/editors earn the amount per hour they want and clients get the productivity gains they're after.
Trados brought about the last major change to the translation industry pricing model; why shouldn't Proz.com and its members lead creation of a new one? ▲ Collapse
| | | | | Pages in topic: < [1 2 3 4 5 6 7 8 9 10] > | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Clients / large translation companies now talking about pooling linguistic data. Should we be there? | 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 » |
| | Trados Business Manager Lite | Create customer quotes and invoices from within Trados Studio
Trados Business Manager Lite helps to simplify and speed up some of the daily tasks, such as invoicing and reporting, associated with running your freelance translation business.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
