Aug 6, 2011

"The Future is here" and the end is near

On September 9th, the Dutch Association of Translation Agencies (ATA the Lesser) will be hosting a conference on the bubbly, bright future of MT post-editors and why all good translators should be eager to hop on that gravy train and ride it to the greatest challenge of their professional careers: turning pig shit into gold.

The conference keynote speech will be by industry prophet Renato Beninatto, which will undoubtedly be full of entertaining claims and predictions. Most of the presentations will be in English, hopefully machine translated to convey the real quality of that present future. Presentations will include a sales workshop by Renato, and Atril, SDL and Plunet will present their products, presumably with some MT-related spin. The rest of the workshop titles are clearly focused on MT editing processes.

Information on the conference program is available in Dutch and English. Those who read both pages will note the date discrepancy on early bird rates. I presume that the Dutch information citing August 10th is correct and someone simply botched the translation and editing of the English page.

Those who need to collect PE points to maintain their Dutch certifications will receive 5 points for attendance.

See you in the future!


  1. Yeah, I'm sure we'll all be clamoring to be there... not!

  2. I disagree, Jill. In a thread about this event and MT in general in a private, off-the-radar forum for translators, one colleague made the excellent point that we need to be present for these discussions. Sort of like Germany in 1933 I suppose. Silence only gives the wrong people the wrong idea that they should proceed with their bad business. I am looking for critter care arrangements so I can attend and be "educated" regarding the latest twists in this path. And I'd love to have Miguel Llorens there and debrief with him afterward with a good bottle of Scotch. It could be fun.

    Also, aside from the MT perversions, I'm genuinely interested in a look at the latest SDL and Atril technologies, and although I disagree with most of his positions on technology, I very much look forward to hearing Renato speak on sales, because it's something he's very, very good at. I've seen few approaches I feel are on the level of what Apple taught its education sales consultants in the late 80s and early 90s, though one does occasionally find useful ideas from the unlikeliest sources. Plus, ya gotta see the pitch if you're gonna hit it.

  3. Well, I'll leave it to you and Miguel to stand our ground then. An MT conference isn't worth me shelling out money on an international plane ticket.

    I've heard Renato speak before. Enjoy.

  4. Hey guys,

    With all due respect, I found your views against MT very hard core. One thing is that the technology might not be as brilliant as some present them to be; another is to absolutely dismiss it.

    With the right implementation, customized MT can be a good tool for translators. My 2 cents as a translator who has been in the industry for 10 years and has tried several MT systems.

  5. Hey Rubén,

    The technology's not the problem - it's not like MT was invented yesterday. it's been around for quite some time, so there's been loads of time to perfect the technology, the algorithms and whatever needs to be done at that level. It's language itself that's the problem, which is why the technology simply can't fly. As a translator, you should know that there is more to using and understanding language than vocabulary and grammar. You should also know that language evolves and changes every time someone opens their mouth to say something. It's just not as straight forward as the MT guys would like us to believe.

    I am personally not opposed to anything that will make my life as a translator and editor easier. Quite to the contrary. But I am not looking forward to customers asking me to check texts 'for grammar and spelling mistakes' that they ran through some MT they found somewhere and translated into a language that is foreign to them. Take into account that most of the people who put pen to paper can't write for the life of them - and this has nothing to do with literacy or lack thereof - I simply do not comprehend how anyone can expect MT to produce anything usable. A poorly written source text produces a poorly written target text. Basta. Human translators can turn shit into a diamond of whichever quality they want. MT will never be able to do that. Hell, MT can't even produce a usable translation from a properly written source text!

    So while everyone professes the demise of the translator as we know him today, and tries to turn bad translators into bad editors, and drive good translators who don't or can't edit out of business, I will make a point of going about business as usual and professing that MT will not, and simply cannot, live up to the promises being made by the people who are trying to make a quick buck. Because that's exactly how I see it: translation is something very few people understand, and the longer it stays that way, the longer and the more some people will be able to fill their pockets. Compare it to the implementation of one of the biggest business application providers in the world and see how that works for you: functionality looks good, budget is spot on, but 10 years down the line, oh no! the program is still no implemented properly because customization never stops and the budget? Oh well, we don't even know where that is at these days - shot through the roof is probably the right answer.

  6. Susan van den EndeAugust 09, 2011 2:12 PM

    Well, I was at last year's ATA conference (TMT: the road ahead) and rather appreciated the "food for thought". Presentations included case studies and some explanations on the differences between free MT (Babelfish, Google Translate) and customised engines that are specifically programmed for a certain project.

    I for one concluded that with the present technology, MT with a customised engine can be a good choice in some very specific cases. But if you want to do it right (this customisation isn't free), the break-even point probably lies at 70-100k words or so.

    No-one from the "pro-MT" side at last year's conference claimed that MT would work for small files or badly written files. They are the first to point out that MT can only be useful if you meet a whole set of specific criteria. But, as Rubén said, MT *can* be a good tool for specific situations, with the right implementation and customisation.

    If you think that MT equals Google Translate, a conference like this one is guaranteed to bring you some new insights. What you do with them is up to you. But don't dismiss MT before you know what exactly you are talking about, and most of all, don't dismiss colleagues who choose to work with MT.

    All too often, the attitude I see is "MT equals bad quality". I don't have MT experience myself, but have seen the output of some MT projects and have nothing but respect for the good work my colleagues delivered there. The client paid about 70% of what he would have paid for translation, was happy with the quality, and the editors involved were quite happy with their earnings as well.

    Agreed, MT will never work for all projects, so no need to fear for your livelihoods. But it can work for some, and there's no need to be so 'hard core', as Rubén said.

  7. Hey Susan,
    I'm going to this year's conference because I'm interested in hearing about and seeing the latest developments.

    I have some experience with MT, and I have so far not been impressed with the results, not even in terms of the post editing requirements. And the MT guys I have heard so far are not so careful with their promises. I do, however, agree with you and Rubén that MT can be useful in *some* environments, but they are few and far between, and the 'whole set of specific criteria' is not to be viewed lightly and includes everything from the creation of templates to teaching people to write in a specific way, i.e. to use specific words to mean specific things and to put words in a specific order. That is a challenge in itself! But as long as the customers buying these systems realize this and have a business case that proves the investment is worth it, they would be stupid not to do it.

    Like I said, I am not opposed to anything that makes my life as a translator and editor easier, and I am not dismissing colleagues who choose to work with MT. We all have different strengths and preferences, and so we all have our place in the market. What I am opposed to are the sales tactics used to sell an application that is useful in *some* environments as the newest need-to-have gadget to seople who know nothing about translation at all.


Notice to spammers: your locations are being traced and fed to the recreational target list for my new line of chemical weapon drones :-)