Wednesday, July 14, 2010

What is Holding the Wider Adoption of Machine Translation (MT) back?

I have been seriously distracted the last few weeks by the World Cup which is really the only sporting event I really ever connect to anymore. This year I watched more games (in HDTV too) than I have since my childhood and I am glad that the clearly better team won the final. My opinion on the real winners: Germany, Uruguay, Spain and Ghana, the biggest and possibly worst loser in World Cup history: The Netherlands. 

Anyway, there is an interesting discussion in the LinkedIn MT forum started by Lori Thicke that I thought was worth summarizing and highlighting. The original question she poses is: Why is MT adoption so slow, given the clear and real benefits that she has seen in her own experience with MT? I would recommend going through the whole discussion as I am only selectively paraphrasing some of the comments (IMO the best and most informative) and the comments are worth reading in their original state. This is my summary which is in no particular order.

The Reasons Behind Slow Adoption

  • Translator Resistance (could it be because many have direct experience with low quality over-hyped MT or insufficient training?).
  • Lack of Training on how and when to use MT.
  • People who think that MT is a replacement for human translators (especially those who expect the same quality as humans). Consumers need to understand that MT + Human is the construct necessary to get the best quality.
  • Too many eMpTY Promises from MT vendors who claim to have the silver bullet and deliver mud. 
  • There is a lot of low quality and erroneous advice (mostly opinion) that is outdated and obsolete on translator bulletin boards and forums that malign and present MT as a threat or a crappy technology.
  • Very few understand that customized MT engines are quite different from the free online engines at Google, Microsoft and Babelfish.
  • MT has been complex and expensive.
  • The poor quality of communication between MT vendors (many who feel offended by the quality criticism) and translators (who feel threatened by lower costs and bad quality and reduction of their value and role).  Marco Cevoli has provided a presentation on this conflict here from a slide deck presented at a ProZ conference.
Bob Donaldson points out the disconnect between the vendors and the clients in the translation production chain (Buyer > LSP > Translator) and points to three reasons for this disconnect:
  1. Different understanding and expectations of translation quality through the chain
  2. Lack of Translator (Vendor) Acceptance because of poor project management and skill mismatches in the editing and cleanup (post-editing) process.
  3. Inability to accept risk and manage it in a way to meet customer objectives, and without expecting the translators to bear the burden of poor quality MT systems. LSPs should be willing to share benefits as well as risk with translators as MT systems improve and ensure that they ACTUALLY raise productivity.
  • Translators may also not be the best post-editors and users will need to find how to find the right balance here. Sometimes less skilled people would be a much better fit for post-editing tasks. The world where MT is commonplace will have many more linguistic roles for people with linguistic skills.
  • I think that better definitions of deliverable MT linguistic quality would also go a long way to making the technology more widely accepted.
  • Many people do not realize that MT + Human Post-editing only makes sense when there is an actual measurable productivity advantage. Don’t use MT if this is not true and getting an MT engine to this state requires investment, effort and know-how.
Joachim Van den Bogaert also made some interesting points:
"Using MT and post-editing requires a paradigm shift. We see that, for now, translation jobs primarily come in project packages: the typical cycle is: engineering (pre-processing) -> translation -> engineering (post-processing) -> QA -> corrections -> QA -> final delivery. MT is at its most efficient in continuous projects: add content -> MT (with included automatic processing) -> post-editing -> delivery. Instead of focusing on delivering localized content at a specific point in time, MT allows you to consider having ALL your content localized ALL the time. Indeed, a well implemented MT chain, eliminates initialization costs, which allows you to have minor changes and additions to be translated instantly. But this requires a different mind set from (traditional) localization consumers."
  • Open source MT like Moses is difficult to use productively without a set of other tools (segmentation, alignment, corpus cleaning and analysis etc..).
  • MT systems have very different requirements and controls and thus it is difficult to integrate with production systems and have common and consistent pre and post processing.
  • In terms of mindset, I feel that MT should be considered as moving to a production line where some things can be built more efficiently as opposed to the craftsman approach that many have today.
Just because 90% of the professional translation industry makes disparaging remarks about MT on a regular basis does not make the technology irrelevant or unusable. Change often comes from the few, who solve many of the problems we have listed above and develop the new models for success. There are now a number of firms in the professional translation world who “get it” and they are likely to pave the way and build competitive advantage through informed and intelligent use of MT technology.

Alon Lavie points out that MT is the only way that new high-value dynamic content like customer support and user generated content (UGC) can be done, I believe this can be done at much higher quality levels when done in partnership with LSPs who are informed about the linguistic strategies to improve and refine MT engines. I have written about this type of content that presents a major opportunity for the professional translation industry. (I think so anyway). However, this will require new mindsets and new translation production strategies. We are in the early stages of this today.

So while many of these problems and issues remain unresolved, there are more and more initiatives that are attempting to address at least some of these problems. We are seeing much these days about how the translation production chain needs to evolve to handle the more dynamic and continuous flow of content that is increasingly valuable to translate. MT will be critical here.

It is clear that we should attempt to address all or as many of these problems listed above as possible but a few stand out for me.

  •  More education and more information about the current MT technology 
  • More information to help users assess where it makes sense and where it does not. 
  • Direct contact and hands on experience with a customized MT system. 
  • Better integration with the other translation production infrastructure. 
  • Development of win-win scenarios in MT + Post-editing work processes and compensation so that all participants share in the benefits of improved production capacity.
So while there are many difficulties there are also many new opportunities that provide new vistas for professional translation market leaders. And while it is clear that MT on it’s own is not the solution for every translation problem, it is also increasingly clear that it is a technology that will have a long-term role in making large volumes of high-value content multilingual at a cost and turnaround time that is viable and furthers business objectives.

And a parting note: for those of you think that MT is not a mainstream technology yet, consider how many people are using free online MT engines each and every day. I don’t have accurate data (please let me know if you do), but I estimate that anywhere from 10 –20 million people translate a web page, a document, a phrase or a word every single day in hundreds of language combinations. (About 2.5B to 5B words daily!) I think we can all safely assume that the thirst for information and knowledge in the growing human online population, has an energy and momentum that is far stronger than the inertia in the professional translation industry. It will drive this technology forward and large volumes of content will get translated at amazingly high quality levels. We are already seeing examples of this.

My advice to you: Wake up and smell the coffee (as they say in America) because often "good enough" is good enough. And I have to again use the 
quote that June Cohen, Executive Producer and Curator of the amazing TED Talks made at SXSW this year (even though it may seem naive and idealistic to some) when she was asked "What technology would you like invented? Or uninvented?"

"Instantaneous, accurate translation online. Nothing would do more to promote peace on this planet."


  1. I don't think translators may not be the best post-editors, even though they could be a major problem for their education.
    I think the problem is the approach to translation that is definitely not economic, and this is fatal because most translation educators are not translators in fact; they are mostly linguists, sholars and devotees of literature. This is a problem when it comes to (translation) industry.
    On this basis, I don't even think that more education and information about MT technology could help. For the sake of their careers, translation educators will keep on maligning...
    for a long time, until they will eventually change their perspective to a more economic one that makes them aware of the actual, measurable productivity advantage. And this will come from the customers, from those who pay for translation, for the product, not for the "hermeneutic process".

  2. I don't know how things are with other translation machines, but I had some insider information from Google and it was not good, because they told me Google does not work with linguists, and does not want to have linguists working for them. This makes things a lot worse. I think we would give them valuable contributions. Engineers can develop wonderful software, but they are not linguists, they cannot judge what needs improvements.

  3. It's a thought-provoking question, because it can be interpreted in different ways.

    1. Perhaps you mean to ask: "Why is machine translation not being used to produce high-quality finished translations?"

    Because it makes too many mistakes! Seriously, that's the main problem. If the results of machine translation could be trusted, it would be more widely used.

    2. On the other hand, depending on one's definition of "machine translation", the premise of the question might be wrong. Machine Translation is in fact very widely used. For example, I use a "translation memory" program. Another name for it is, "example-based machine translation". By this definition, thousands of translators around the worls already use machine translation. In addition, millions of people use online translations from Google and elsewhere, every day.

    3. Maybe your real question is: "I sell a machine translation system, but why don't more people buy it?"

    Many translators and clients don't see the true value of machine translation and how translation can be improved by intelligent use of computers. Good translations can be produced without computer. You don't have to buy a program and learn to use it, or connect to the Internet.

    Part of it is that translation, for many translators and clients, is not a full time job. If you only dig one small hole every month, you will be content to use a garden spade. On the other hand, your job requires you to dig a trench 100 meters long every day, you might purchase a power shovel, mounted on a tractor driven by a gasoline engine. But that requires money, and education about the availability of power shovels, and a factory to buy it from, and training for the person who operates, and a mechanic to repair it, and a supply of gasoline ... It's not a simple or obvious improvement.

    Posted by Steven Marzuola

  4. I can not speak about machine translations other than those between Japanese and English, and the ones I have seen so far have been simply terrible. If a Japanese word has more than one possible meaning, the machine always seems to pick the wrong one for the context. At the current juncture, I think a J-E translation machine would work well only in translating lists of items--not sentences requiring a noun and a verb, not to mention changes of tenses, modifiers, etc.
    Posted by Lora Sharnoff

  5. Languag itself. Machine translation is unable to deal with the limitless adaptabilty of language, which is used differently every time in every situation. Even in standard documents that use standardised technical language, which is the main field of applicability of MT, there are always new forms MT cannot deal with; to say nothing of errors in the source text.
    Posted by Tom in London

  6. From an individual translator's point of view, the inter-related aspects of price, availability and ROI are the drawbacks.

    For a freelancer, buying a TM program is an obvious step. It may be a relatively expensive purchase, but you get a package that you can install and start using right away without putting anything in first. You get your deliverable just as fast, or faster, than you would if you weren't using it. And your ROI starts to snowball in front of your eyes as you work. Pretty much anyone can see the benefits. Getting to use TM for the first time was a joyous experience.

    With MT however, the decision to invest is not nearly so clear. What packages there are don't seem to promise 'plug and play' usability. If it's an open source program like Moses, the initial barrier is much higher. There's no guarantee of when training will reach the point that you can get a deliverable out of the system. The hardware requirements are also daunting for an individual. And so with MT, the ROI seems far more murky.

    Within the ecosystem of end-client, LSP and freelancer, TM fits comfortably in each ecological niche, and uptake is pretty much assured. But with MT, the path to uptake for the freelancer is strewn with obstacles, even without taking into account their fear that MT might replace them entirely or in part. Freelancers won't be very keen on tools where only the big boys hold the levers.

    One of my clients, an LSP, sent me a questionnaire the other day about integrating me into the MT ecosystem. It went something like this; "Have you heard of MT?", "Are you hostile to MT?", "Would you consider doing post-editing?". While I answered yes, no, and yes respectively, it isn't a very appealing approach to the issues. There was no invitation to help design or shape the system, and since LSPs' implementation of TM has been sub-optimal, there's no reason to suppose they'll do a better job with MT.

    So I would suggest that one of the things holding MT back is that it doesn't fit into every niche in the ecosystem yet. Making it fit will require some thoughtful outreach from the companies that use it to freelancers and translators who are open to the possibilities. I don't see that happening at the moment. (Asia Online doesn't even respond to queries from freelancers...)
    Posted by Rod Walters

  7. Hi Kirti,

    Following text was translated from French to English using a MT tool and shows I think why MT is not more widely adopted. Have fun (-:

    "BRUSSELS - can no longer provoke Belgian oud-vorstin the Fabiola (82), thus of its keepers them to hear has got. During national holiday previous year swaid Fabiola (r) defying with an apple photograph: AFP Please Wait… The very catholic woman gets regularly notes in which are announced that she is shot, including with a hanging-post truss. During the national defilé previous year she swaid defying with an apple, referring to the tale of Willem Tell. „The security services were not possible there absolute to laugh. For this reason the queen was now asked to this way no more jokes remove, weet the newspaper to Gazet of Antwerp communicate Monday. Fabiola are the woman of king Boudewijn, who governed Belgium between 1960 up to 1993. She decreases a defilé of especially soldiers to Wednesday with its family members as from a podium in Brussels again. Reason is national holiday, on which Belgium celebrates that it got its own king after the dissidence of the Netherlands. Fabiola will have refused a thin kogelwerend carry waistcoat. Its keepers carry, however, boekentassen with kogelwerende shields. The access ways to the defilé are turned off with all kinds of obstacles to prevent attacks like on the Dutch royal family in Apeldoorn, previous year."
    Posted by pierre leonard

  8. @Pierre
    Coming with an example of bad MT is a pointless exercise.

    It is clear that the way to proceed with MT in professional environments is to customize it. That is how you get good quality, not by running some probably poor quality source through a free online engine.

    I could just as easily show you very compelling examples of high quality MT output but it seems that you have already drawn your conclusions so we can leave it at that.

    Yes I agree it is very easy to get a bad example of an MT translation.

    1. I am flummoxed by highly touted multinational corporations working through a closely allied big-city agency, who then ask you to become a part of their MT post-editing team and say: "Oh, you have to accept a base per-word fee no matter if the content is poor or very good. Swallow that whole, you low-leverage freelancer, for we refuse to pay you for time spent from a highly skilled professional linguist. (A comedian might say: We would like you to pose as the inhabitant of a salmon farm. Silence on the set, please.)

      And if you raise a legitimate objection, they all go silent (like we fish arer supposed to be as a means to an end).

      See below if you want to prove me wrong - it is not hard to find me using "verharmlosend" as a search.

      I always respond by saying that were I an end customer of a Fortune 500 company, I would be very displeased that the LSP has put an incentive in place for linguists/editors/fixers to "work fast" or "barely work at all" on certain content just so they achieve enough earnings per hour to make a decent living, cover a mortgage, needed holiday, benefits and insurance all paid up, but give the freelancer no control to set his terms, since he is an independent contractor. Instead the measured billable unit is dictated by industry/the agency.

      Really bass-ackwards, if you aks ;-) me.

      In mirth: yesterday, Bing translated (from German to English) "klingt etwas verharmlosend" as

      "sounds somewhat reverse ham antispasmodic"

      A correct rendition would be more like
      "sounds somewhat downplayed (in importance)".

      Now you get offered 15 percent of a translation fee to fix that. And are told to work fast to make enough money.

      Of course MT can amass a lot of subject-specific and accurate results in its parsed sentence-phrase-word pairs -- point granted. But the lack of empathy for what freelancers face is sad, but true, and very disconcerting.

      If there were all sorts of agencies offering a fair arrangement for uncluttering and making sense of MT efforts, maybe after many years in the business, I would have heard something about it.

      I invite you to prove me wrong.

      Or my sources (not the texts) are all flawed.

  9. Hi Kirti,

    To the problems you listed for immediate attention, you can add the following:

    Better understanding of localization/translation by MT practitioners.

    So far most MT practitioners/advocates, with the exception of Jeff Allen, have not shown enough interest in and understanding of the processes, concerns and other issues of the translation business. Worse, the dominant attitude among them is "you need to learn our game", which is witnessed in the list you provided. It is the translators that need to be informed and educated. This attitude in itself alienates the translator community. In contrast, it is much easier to communicate with Jeff Allen on MT and accept his ideas, because he knows how to address the concerns of translators/LSPs and how to convince them. If the MT people know as much about translation and the translation industry, buyin from translators will come much easier.

  10. Hi Kirti,

    I didn't get your point immediately and posted indeed a bad example of MT on purpose.

    As a translation vendor, my point is to demonstrate how many MT solutions give poor results.

    I am very often talking to clients who don't see why they should (quote) "pay for having their website translated in funny languages while they can use Babelfish" and end up with rubbish that ruins their online presence and reputation.

    I had a similar discussion on another forum a few days ago. The question was I think "is translation science or art?". My answer was: it depends on the kind of text you have to translate.

    I agree that various techniques - simple writing, crowdsourcing, statistical MT, TMs - can be combined to get accurate results for texts that don't need any kind of transcreation or cultural adaptation, like technical brochures, user manuals, catalogues of mechanical parts etc.

    (One could argue "but then why are so many user guides so poorly translated?" but that's another story.)

    On the other hand, trying to sink costs at any price is useless. These tools are no good for newsletters, press releases, PR or marketing texts, most online content or any text with colloquials, double meanings etc. - not to mention subtitling, where you have to translate and summarize at the same time.
    Posted by pierre leonard

  11. @Pierre

    Thank you for your clarification. You are of course entitled to hold whatever opinions make sense to you.

    I have however, noticed that most of the MT naysayers have had NO experience with a customized MT engine of any kind. Or they may have had one bad experience.

    MT is coming whether you like it and agree or not. The forces that are driving wider use are outside the professional translation world and often have to do with poverty eradication and fundamental human access to knowledge.

    Ultan provides a very clear explanation in his timely blog entry at:

    Anway, thank you for your clarification.

  12. @Steve The question here is in fact: why is adoption low even though it does work well when used properly? While it may not be widely known there are many very successful MT implementations that prove the technology can work both in standard L10N projects as well as in new applications.

    @Rod unfortunately developing MT, especially SMT is a large user/LSP solution and not well suited for individual users at this point in time. Any MT requires some upfront investment with uncertain benefits a priori and until this becomes more predictable it is likely that this is not easy for freelance involvement. The free online solutions can only give you the sense for what an un-customized engine experience is like. RbMT may be an easier place to start for an individual freelancer.

    @Frank I agree that there needs to be a higher quality dialogue between people in localization and MT vendors. I think this is beginning at Asia Online and some RbMT companies and we will start to see this bear fruit in the near future.

  13. It strikes me that we spend more time telling people how MT should NOT be used, and what it is is NOT, instead of telling people about MT applicability and advantages and what CAN be acheived. That is wrong.

    The very term "machine translation" plays right into the hands of those who cite examples of mangled output where the free online MT tools were let run riot. Just like the term "controlled authoring" (a disastrously negative, threatening term for anyone trying to use when trying convince content developers about information quality tools), perhaps we need to consider how we brand and position the technology on a general level with translators and potential consumers of high value information.

  14. Hi Kirti,

    In your reply to Pierre, I see the familiar attitude again:

    "MT is coming whether you like it and agree or not. The forces that are driving wider use are outside the professional translation world and often have to do with poverty eradication and fundamental human access to knowledge."

    You are wondering why professional translators do not come to MT, and in the meantime you are saying what they think or say does not matter. This attitude is one of the reasons why MT failed to be adopted more widely. I don't see how you can push for MT without the participation of professisonal translators.

    By the way, your argument was true 10 years ago, but L&H PowerTranslator did not get very far despite the need. The crucial point is not the need, but the maturity of the technology. The need is always there. Without the right technology, the need will remain untapped, as it was 10 years ago.
    Posted by Frank Wang

  15. Frank,

    When I say "whether you like it not" I of course mean you, we or anybody. The major driving forces behind the technology are not professional translation companies. Government, Search Cos. and general education initiatives are much more committed to this because they have a problem that cannot be solved any other way.

    The unpredictability of an MT project is a (if not THE) major cause for the lack of adoption. MT projects require investments of time and money with no definite promise of success and ROI. Nobody can really predict the error levels that will be delivered before the MT engine is done. This is a very difficult scenario for most in the L10N industry to deal with. Some MT investments can fail. I think this factor is a much more important factor in holding back adoption, than the inability of MT vendors to present the benefits in a translator friendly format. In the blog I was just listing the reasons many others had expressed. Yet in spite of these problems, there are still many who have engaged with a productive MT application and there are a growing base of successes. You can see this through use case scenarios presented at TAUS, or even just by noticing that SDL thought LW was 2X as valuable as Idiom in $ terms. MT momentum does not come about just because of good sales efforts - there usually needs to be some substance somewhere even if it is not very visible.

    Today we are at a "maturity" level where many large global enterprises see the value of the technology in its current state, and you can see deployments growing. I think it will get easier as more MT projects provide data to make it more predictable and as this experience builds and eventually includes translators. The few translators who engage early (many already are –not just post-editing but in thinking up improvement strategies) will probably lead the way for the many. I think it is most likely that translators will teach other translators on why it makes sense and how to engage in a constructive and economically beneficial way.

    I am engaged with several professional translation firms and though the dialog is somewhat uncomfortable at times, there is definite progress. In the meantime we can only continue to try the best that we can to communicate the benefits and value, even though some may see these efforts as problematic and imperfect. Linguistically skilled professionals will be necessary to steer and drive the best MT engines and hopefully we can all articulate the skill set necessary to be successful quickly.

  16. 1. Lack of understanding of the target scenarios - on users' part and unfortunately often on developers' part as well.

    MT is a different business case of translation; it is not supposed to go where the human translators go (in most cases). There is a reason why McDonald's is not trying to compete with Michelin rated gourmet restaurants. They both serve food, but for different business scenarios. Amazingly, same people may dine in both on different occasions.

    Here comes Pierre and shows that Big Mac (and, BTW, like Kirti mentioned, this Big Mac is below the average) does not taste the same as fois gras :-) .

    2. Lack of business infrastructure.

    How many multinationals have budget for MT and potentially consider MT to cut costs? How many of IT managers understand the subtle divisions between different types of MT?

    3. Lack of business-minded MT people.

    As of today, the MT industry is staffed primarily with enthusiasts. Many of them are coming from academia. Proving a theory or demonstrating an algorithm may not require the same skills as building a good product.

    Building a good product is halfway of providing a service which the customer finds satisfactory (note, I don't say "good" because these normally overlap but not coincide).

    4. Opposition of (how do I call this...) "pro-human" factions. The usual argument is a poor quality of MT.

    There is no denying: just like in other areas, on some point, machines take job opportunities from humans. I remember talking to a freelance technical translator who mentioned that after Babelfish was made public, he lost a significant portion of people who came to translate patent applications in bulk. Do they really need to know the exact content of every one of them? Not really, they just wanted to see that their patent don't overlap with someone else's patent.

    On the other hand, another translator presented a scenario where MT would essentially improve the quality of human translation and the business experience as a whole. When a customer has thousands of pages where they need to find just one item, there are few ways to handle this. One, pay thousands to human translators. Not many can afford that; the project might be simply axed. In the best case scenario, the translators will work day and night rushing to complete all these largely pointless translations. Haste makes waste. Another scenario would be simply running the bulk through MT, and asking the human translators only translate the necessary parts.

    So it's not all black for the human translators.

    What happens when the software evolves further? I guess the market for the human translators will transform rather than just shrink. They will need to be more technically savvy and learn to live with MT.

    Posted by Vadim Berman

  17. Kirti,
    I have followed the debate both on LinkedIn and here. In both cases, nobody seems to talk about terminology management anymore... As many other people, I have seen examples of texts translated with MT, but I have never actually worked with a customized tool. As a translator, but, most of all, as a terminologist my question would be, does terminology management still play a role in MT? And, could you describe that role?
    Thanks for helping me understand more in this respect.
    Laura Di Tullio

  18. Laura

    There actually is some discussion about how clean and consistent terminology can provide a strong and positive impact to the quality of SMT (or any MT) engines in this blog article:

    SMT engines in particular improve with consistent terminology since it is possible to build statistical mass around key phrases (terminology. There is a slightly different perspective as SMT may also consider, all high frequency phrases as Terminology. I think input on terminology will continue to be a critical skill for developing high quality SMT (data driven MT)engines as it evolves and incorporates more linguistic feedback. The study referenced in the link above provides some examples.

  19. very, very interesting!

    I was just few weeks ago a fierce antagonist of MT, but after a test I change entirely my mind

    I used a simple demo, but with it I realized to be able to translate with the same quality in a half time

    Even if I'm still a fan of Slow Work, it's not a contradiction, as if I complete a job in half time, I can spent double time to study and/or doing other tasks (not necessarily job-related)

    Clearly, I think that MT is still useless without a brain behind, i.e. only a sound translator that know VERY well the topic can use a MT to translate

    Anyway, I think I'll use more MT in the future and I am open to MT post-editing too

    tomorrow I'll follow the Allen Proz webinar and in the meanwhile I'm studying all possible material on the matter

    then, thanks to Kirti too!

  20. Hello Kirti,

    I am still digesting what I've seen and heard during the ATA and AMTA conferences in Denver, while trying to visualize a "Man cum Machine" panel discussion at the next ATA conference. Hopefully, some of my first impressions and preliminary thoughts may be of interest to you and your readers.

    (I probably could be classified as a technology-friendly Luddite, i. e. I gladly adopt unnamed bits and pieces of various technologies that fit and enhance my (constantly changing) work processes, but, at the same time, I am very likely to be somewhat skeptical about new "named technologies" and productized solutions.)

    One problem that I can see at this time is that the MT community seems to aim way too high to gain positive interest from smaller LSPs, including freelances.

    Accordingly, I feel that making, for example, tools kits like Language Studio Lite and other useful scraps of your technology more accessible for individual translators and small LSPs as end users, and possibly future loyal paying customers would go a long way towards overcoming HT resistance.

  21. Kirti,

    Thank you for your excellent insight into this issue. I think what is truly holding back a wider adoption of MT is simply time. In "time" the MT Engines will get so adaptable, customizable, quick at learning and user/operator friendly that the adoption barriers will all come down. This is already happening and very quickly. I can understand Translator concern regarding this but not really the anger. It's like being a maker of buggys and being angry about the appearance of the automobile.

  22. @precaryus - Thanks for the comments. I think the translator anger comes from the way in which technology is used to reduce and change their compensation model. I think it is quite possible that if this changes, more people will get involved. The real culprit for this I think is the way TM fuzzy matching percentages are used to commoditize and reduce rates paid to translators. Quite often this is done in an disingenuous way. Hopefully early MT users will compensate editors fairly for their work and establish models that nurture all the players in the eco-system.

    1. "Hopefully early MT users will compensate editors fairly for their work and establish models that nurture all the players in the eco-system."

      Based on my experience from approaches by agencies in Europe, the ones I know are not adopting this kind of model. (Or I would have heard about it.)