Pages in topic:   < [1 2 3 4 5 6 7] >
'[MT] is most often used alongside [TM] as an adjunct to human translation'. Are you using it?
Thread poster: Henry Dotterer

Samuel Murray  Identity Verified
Netherlands
Local time: 12:34
Member (2006)
English to Afrikaans
+ ...
SITE LOCALIZER
Sorry if I misunderstood Sep 2, 2009

Didier Briel wrote:
Samuel Murray wrote:
Didier Briel wrote:
Have you checked Autshumato ITE? ... It ... includes also machine translation.

Well, you can think for yourself... machine translation between 11 languages, packed into 7 megabytes?

I didn't say everything was packaged, ready to use. By "check", I mean try and get information.


Sure, sure. No offense intended. You said "check", so I checked, and now I'm reporting back on what I've checked. Also my apologies for having misunderstood your comment that it "includes" machine translation. The product's blurb mentions machine translation, but the user manual doesn't.


 

Jeff Allen  Identity Verified
France
Local time: 12:34
Multiplelanguages
+ ...
MT for Afrikaans & how much data for SBMT Sep 2, 2009

Jeff Allen wrote:...I know that there has been a lot of interest in over the years for MT systems for languages in South Africa, because there are several language there which have official status. And such official status often generates a need for translation.


Samuel Murray wrote:
I'm not aware of any MT systems being developed for (non-Germanic) South African languages. I've heard rumours of an English-Afrikaans engine developed by some university about ten years ago, but although I encounter the rumour time and again, I've been unable to confirm it myself.


It was the:
Unit for Language Facilitation and Empowerment
University of the Free State

I met a person from that dept about 10 years ago (so your time period estimation corresponds) at a translation/localization conf where I gave a presentation and we corresponded for about a year. However, I never saw any result, and I think the contact's email started bouncing several years ago.

On topic of MT and TM compatability:

Samuel Murray wrote:
If my CAT tool claims to work with Google Translate, and something doesn't work, I can turn to the CAT tool vendor and expect him to sort it out. But if I use a commercial MT system, and my CAT tool fails to connect to it, my CAT tool vendor (who does not own a copy of each and every MT system on the market) might have cause to say "well, ask the MT vendor to sort out the problem" (and the MT vendor will say a similar thing about the CAT tool vendor).


You have hit on an important point that goes up and down over the years. They to handle it is that there are only a couple of MT systems that have true TM plug-ins.
And several of the MT systems now have internal TM-type subsystems, so you can convert your TM tool TMs into TMX and then republish that with the MT system's internal TM, or attach the TM as a pre-processor, similar to as how custom dictionaries are attached and override the processing.

You will notice that for the time that both MT tools and TM CAT tools were both focused on processing in MS Word, the TM tools were using that format as the means to do the MT processing along with the TM. But now that many of TM tools are expanding out to other native formats, and the MT tools are also going to other types of document and content formats, the compatability between the two types of commercialized systems has become less and less. Yet, the Google API makes it possible for several of the TMs to use that way to tap into an MT systems. However, again, they are pulling the MT data processing from non-customizable MT engines, so the quality level can vary from domain to domain, and not way to feed customer specific terminology into it.

And as for how much data......

Jeff Allen wrote:
Well, there is the possibility of taking an opensource system and training it with your own data, but that just not really possible to get anything usable.


Samuel Murray wrote:
Do you happen to know how much data is necessary to train an opensource system up to a useable level (assuming we can define "usable level" easily)? I realise it is a silly question (with relativity and all that) but I'm hoping for a vague indication here.


This is difficult to answer, and I asked the same question a couple of months ago, and a Statistics MT vendor replied on a list as follows below in italics.

My experience includes statistical MT, but in a context of creating hybrid MT systems. However I should add that the reply below is from a vendor that spends 100% of their time working on this, and is constantly investigating how to help corporate clients aim at having consolidated, clean data. This means that they have to create scripts to clean up the TMs and the other data, which was stated as taking 50%-65% of the project time. So it is by no means a small task, and they focus on corporate/enterprise users. Such projects also focus on building separate statistical systems per content type or subdomain. So, these are not little projects at all, and the objective is enterprise-wide translation.

"The question of how much data is needed is difficult answer with a single number. As a general rule , the more tightly focused the domain the less data required, and the more general and broad the domain the more data required. Data volume required for success can also vary by the language pair. Good FR > SP systems can be built with as little as 10K TUs. Unlike language pairs need more data. Generally, 200K+ is what you need for FIGS systems to get to a level where post-editing is fast and productive. Glossaries and large amounts of high quality monolingual data that can be easily harvested off the web are also very useful linguistic data that can help drive improved quality.

These are very, very rough generalizations to provide a sense of proportion.

The time involved in SMT engine development is linked directly to the quality one wants to reach. As a general rule, more time means better data and better engines. If you don’t know where your data is then you need to get a handle on that first.
....

Ongoing post editing corrections can be fed back into the training system when they reach a certain mass, say 1,000+ corrections to retrain the engine so the engine continues to improve on a weekly basis.

....

The long-term ROI and TCO of SMT systems is very compelling for people who are considering EXPANDING the scope of what they are going to translate. SMT is a strategic change for a localization department and is also a process of evolution... "


[Edited at 2009-09-02 21:49 GMT]

[Edited at 2009-09-02 22:00 GMT]


 

Jeff Allen  Identity Verified
France
Local time: 12:34
Multiplelanguages
+ ...
MT postediting and the EC/EU Sep 3, 2009

Eric Hahn wrote:

"The calls are generally published in the Official Journal of the European Union, C series and may relate not just to translation proper but also to the rewriting, modification and revision of texts and the _rapid post-editing of machine translation output_."

Source : http://europa.eu/languages/en/chapter/35



The European Commission has been issuing calls for external Posteditors since 1998. See footnote 3 in:

ALLEN, Jeffrey. 2003. Post-editing (chapter 16, pages 297-317). In Computers and Translation: A Translators Guide. Benjamins Translation Library, 35. Amsterdam: John Benjamins.
http://www.geocities.com/mtpostediting/published-PE-book-chapter16-jeff-allen.htm

I knew of a few agencies participating as such posteditors in the late 90s and early 2000 years. the EU MT system is a very specific version (evolved version of 1976 Systran system with changes made by internal EC/EU teams). Content is simply processed and sent out in batches, and I have not seen that external posteditors have the possibility to feed back into the system. So it is limited in customization for such external
postediting participants.

Jeff


 

Jeff Allen  Identity Verified
France
Local time: 12:34
Multiplelanguages
+ ...
are the several subtopic discussions OK in this thread? Sep 3, 2009

Hi Henry,
In your original post of this thread,
Henry D wrote:
I wonder: how many of you are using MT now, alongside TM, in your work?


You gave the following guideline,
Henry D wrote:
I hope to get feedback from a variety of people. Just in case this thread might be contentious, I would ask in advance that you limit yourself to one or two posts.


Several of us (me included) seem to be going over the 1-2 posts per head, but not for reasons of highly contentious. Rather, I hope that this exchange is providing a lot of benefit to the audience at large as well as the participants.
There are good specific questions and some answers. Do you feel the questions/answers are hijacking the original point of the thread, or is this OK?

It seems like it would be hard to reshift specific topics over to dispersed threads elsewhere. But if you think we need to migrate some discussion topics over to the threads where it was already discussed, please let us know.

I can say that several questions are providing good grounds for some fine-tuning of of recommendations of MT use, due to specific language pairs, system constraints, software compatability, etc. These are the types of points that MT vendors need to constantly keep in mind with regard to end users.

Jeff


 

Jeff Allen  Identity Verified
France
Local time: 12:34
Multiplelanguages
+ ...
new presentation on En - Hindi MT Sep 19, 2009

Jeff Allen wrote:

I had forgotten about the following post that I made a couple of years ago on the status of availability of HI MT systems.

Eng>Hindi MT
http://www.proz.com/forum/translator_resources/17428-englishhindi_mt:_does_it_even_exist.html#174630




See post I just made at:
http://www.proz.com/post/1219103#1219103

proceedings of MT Summit XII (2009) (Ottawa, Canada)
http://www.mt-archive.info/MTS-2009-TOC.htm
presentation by:
R.M.K. Sinha
Indian National Translation Mission: Need for Integrating Human-Machine Translation

http://www.mt-archive.info/MTS-2009-Sinha-1.pdf


 

gianfranco  Identity Verified
Brazil
Local time: 07:34
Member (2001)
English to Italian
+ ...
MT application field (very large projects developed in a long term life cycle) Sep 19, 2009

From my experience using a commercial product, a PRO version with customizable glossaries, MT technology can be used probably with some long term benefit only if the following conditions are met:

  • 1. the project is very large, in the range of several million words

  • 2. the source text is controlled in syntax and terminology, uniform in its format and quite repetitive

  • 3. there are no significant occurrences of idioms or special constructions in t... See more
  • From my experience using a commercial product, a PRO version with customizable glossaries, MT technology can be used probably with some long term benefit only if the following conditions are met:

  • 1. the project is very large, in the range of several million words

  • 2. the source text is controlled in syntax and terminology, uniform in its format and quite repetitive

  • 3. there are no significant occurrences of idioms or special constructions in the source text

  • 4. the desired output is not concerned with style or any other type of linguistic embellishment

  • 5. there is a team (*) of people preparing adequate reference material, aligned corpora, terminology database, etc. and to pre-edit the source text, where necessary

  • 6. there is a team (*) of people to review and post-edit the text after MT processing (for each target language)

  • 7. it is acceptable an initial costs to setup the system that may be higher than a normal translation, in exchange for future advantages when the project is mature and running


    (*) Note: I'm talking about 'teams', and not single editors or single reviewers, due to the size of the projects (see point 1.)



    Failing any of the conditions listed above (and probably all of them must be met!!) a workflow including MT is unrealistic it doesn't provide any advantage in speed, let alone in quality, or would not be successful. Trying to convince freelance translators otherwise, and not spelling out all the real conditions necessary to use this technology, is misleading.

    Examples of useful applications are those cases where a manufacturer of some product, or service, produces documentation running into several million words, with significant parts of it duplicated across various products, written using a controlled style and standardized terminology, also recurrent across various product, and having a high degree of control on the source text. This may happen in large industries such as automotive, electronics, and similar, but not in most of the daily projects that a freelancer faces every day.

    A workflow that includes MT is *not* useful for smaller projects, (below 500,000 words), not useful for standard freelance work, variable subjects, documents in multiple formats, uncontrolled input language, and, even more relevant, not useful for requests of high quality output.

    In all these cases, any MT output requires as much work, or even more, than a normal workflow, which nowadays includes the possibility of using a Translation Memory (where applicable) and some kind of terminology database as a reference. And this is more than sufficient.

    Gianfranco


    [Edited at 2009-09-19 17:13 GMT] ▲ Collapse


  •  

    Jeff Allen  Identity Verified
    France
    Local time: 12:34
    Multiplelanguages
    + ...
    not all these points are required for MT Sep 23, 2009

    gianfranco wrote:

    From my experience using a commercial product, a PRO version with customizable glossaries, MT technology can be used probably with some long term benefit only if the following conditions are met:

    .....
    Failing any of the conditions listed above (and probably all of them must be met!!) a workflow including MT is unrealistic it doesn't provide any advantage in speed, let alone in quality, or would not be successful. Trying to convince freelance translators otherwise, and not spelling out all the real conditions necessary to use this technology, is misleading.

    Examples of useful applications are those cases where a manufacturer of some product, or service, produces documentation running into several million words, with significant parts of it duplicated across various products, written using a controlled style and standardized terminology, also recurrent across various product, and having a high degree of control on the source text. This may happen in large industries such as automotive, electronics, and similar, but not in most of the daily projects that a freelancer faces every day.

    A workflow that includes MT is *not* useful for smaller projects, (below 500,000 words), not useful for standard freelance work, variable subjects, documents in multiple formats, uncontrolled input language, and, even more relevant, not useful for requests of high quality output.



    Gianfranco,

    From what you indicate, your experience is limited to 1 specific MT system, and only in large enterprise context.

    All the items you listed in your post above are not required for an MT implementation. They can help, and are especially important for such large complex projects, but they are not necessary for any and every type of MT implementation project.

    I have worked on various projects over time with a total of about 10 different MT systems (and sometimes working with a range of basic to advanced versions of specific brands). Each of them has their differences, and it is possible to do similar tasks with the various systems, but in different ways.

    1. there are various commerical MT tools which can in fact handle the customizing of idiomatic expressions, including features to handle maximum word distance

    2. controlled language source text is not necessary. it is indeed helpful, and I spent a few years in that area, including having organized at 3 of the 5 international workshops on Controlled Language (and presenting papers at them). Most of my work since 2003 has focused on not working with controlled source text, because this corresponds more closely to what the translation community faces more frequently.

    3. introducing stylistic variation is indeed possible with MT. I've done it on projects. It all depends on which MT system you are using, and your mastery of all the features of the tool. It's simply the fact that few people performing MT are interested in introducing stylistic variation into the translation. But this does not mean it cannot be done. It can.

    4. There is nothing unrealistic nor misleading in telling freelance translators that they can use MT and use if successfully. Many of the projects I have published articles on during the past 6 years were specifically putting myself in the context and shoes of a freelance translator who would only have access to the MT tool (desktop version) and MS Office products. And on one of the projects, I was doing some of the tasks standing up in the subway, working with a pocketPC in Pocket Excel and Pocket Word. All of the tasks, time, etc are clearly logged in the paper describing that project (and all of the translated output is provided as well). Freelancers can use MT systems. It's just that none of the MT providers has focused on providing training to make the professional translator community successful with the tools. And the few translation agencies that have figured it out, have often strugged to see how to make it useful for them, due to the lack of such training. So, they often try to only pigeon-hole MT into specific big project workflows.

    5. Speed. Well, all the speed and productivity statistics are clearly logged in each of those papers, and in much detail. And it is clearly faster than the stated industry average of 2400 words per day (without use of TM tools). And I also conducted a survey in 2004 on translation speed and wrote one of my articles which harmonized all of the results from over 60 respondents from translation agences, depts, and freelancers to obtain such a single figure.

    6. Translation volume. All of the projects that I have conducted and written these articles on are specifically less than 500,000 word projects. Some of them are just a couple of thousands words.

    7. Cost. It is possible to start with a basic home version at 50$ US and customize the dictionary entries in novice mode on a couple of different MT systems. The features are limited and it is not possible to do advanced part-of-speech dictionary entries, but 50$ doesn't sound that costly to me. As for professional/expert versions, they have more features and range from 500-1500$, and some provide some very advanced features. But the key is being trained. It is better to buy a solution for 500$ and ask for 1-2 days of intense training on it, and then you will have much more opportunity for success than if you try to figure out the dictionary modules by yourself.

    8. MT cannot be used for high quality output. Well, I don't know if you can say that to the companies that I used MT with to produce their marketing brochures, their CD boxes, and their user guides. And they had native speakers check the target translations. Very little changes were made. And they were completed much faster than the same materials translated into other target languages not using MT.

    9. MT not for variable subjects.
    Well, I've used MT with telecom, general computers and IT, automotive, heavy-machinery, job announcements, legal HR docs, adoption documents, legal statutes of an association, religion/theology, electronics, health and education docs, and many more topics. Some were marketing materials, press release, customer acceptance test plan, others were technical docs, others were HR related, others were legal docs, etc.

    The key here is could I just give an MT software program to any freelance translator and expect them to achieve the same level of implementation with no help.
    Not at all. Just the same as it would be difficult to expect anyone to create multidimensional slicing/dicing analysis charts with MS Excel without taking an advanced course on the topic.

    Most of my posts here on ProZ over the past couple of months provide links to evidence of all of these items, and in several cases, I provide links to the content (source and target) when I was allow to share it.
    Yet, if there are any objections to anything written in any of those papers/articles, I'm all ears to what is inaccurate, impossible, etc.
    A few professional translators have taken the time to go through all of those papers and many of my posts, and have been able to use MT more efficiently.

    Hope that helps.

    Jeff


     

    gianfranco  Identity Verified
    Brazil
    Local time: 07:34
    Member (2001)
    English to Italian
    + ...
    Titivillus at work Oct 6, 2009

    Dear Jeff,

    for some reason, MT makes me think of Titivillus, the demon in charge for collecting all mistakes made when we write or talk, including mispronounced, mumbled or skipped words.
    All the mistakes will be taken to Hell and counted against the offender.

    See http://en.wikipedia.org/wiki/Titivillus
    "I am a poure dyuel, and my name ys Tytyvyllus ... I muste eche day ... brynge my master a thousande pokes full of faylynges, and of neglygences in syllables and wordes."


    In the past, Titivillus was always ready to introduce mistakes in the manuscripts, when the monks lost concentration for a moment. Clearly the productivity was rather low, but today, with computers and MT, he has a chance not only to increase the productivity but perhaps even to get a promotion.

    In my opinion, Machine Translation technology is exactly that: Titivillus' way for increasing its productivity and bring a lot more mistakes to his master Belfagor.

    Gianfranco


    PS: edited to fix some mistakes introduced by Titivillus


    [Edited at 2009-10-07 11:08 GMT]


     

    Tim Drayton  Identity Verified
    Cyprus
    Local time: 13:34
    Turkish to English
    + ...
    French>English v. Turkish>English Oct 7, 2009

    My thesis is that the quality of the results produced by the statistical method employed by Google Translate will vary in direct proportion to the similarity of the syntactic structure of the two languages involved. It is clear that with two similar languages you will find many instances of clusters of words in both the source and target text which correspond in meaning. It thus makes sense to search for statistical correlations between clusters of words in the source text and in a large sample ... See more
    My thesis is that the quality of the results produced by the statistical method employed by Google Translate will vary in direct proportion to the similarity of the syntactic structure of the two languages involved. It is clear that with two similar languages you will find many instances of clusters of words in both the source and target text which correspond in meaning. It thus makes sense to search for statistical correlations between clusters of words in the source text and in a large sample of target language texts. Where this correspondence does not exist, as with languages from different families, I think that the statistcial method, unless it is developed to take differences of syntax into account, will always fall flat on its face.

    By way of example, article 18 of the Universal declaration of human rights reads in English:

    "Everyone has the right to freedom of thought, conscience and religion; this right includes freedom to change his religion or belief, and freedom, either alone or in community with others and in public or private, to manifest his religion or belief in teaching, practice, worship and observance."

    and in the official French version:

    "Toute personne a droit à la liberté de pensée, de conscience et de religion ; ce droit implique la liberté de changer de religion ou de conviction ainsi que la liberté de manifester sa religion ou sa conviction seule ou en commun, tant en public qu'en privé, par l'enseignement, les pratiques, le culte et l'accomplissement des rites."

    A highly accurate human Turkish translation of this same article reads:

    "Her şahsın, fikir, vicdan ve din hürriyetine hakkı vardır; bu hak, din veya kanaat değiştirmek hürriyeti, dinini veya kanaatini tek başına veya topluca, açık olarak veya özel surette, öğretim, tatbikat, ibadet ve ayinlerle izhar etmek hürriyetini içerir."

    A comparison of Google Translate's translations of the two versions of this article, I think, make my point abundantly clear:

    Google Translate's translation into English of the French text:

    "Everyone has the right to freedom of thought, conscience and religion; this right includes freedom to change religion or belief and freedom to manifest religion or belief either alone or in community with others and in public private, teaching, practice, worship and observance."

    Google Translate's translation into English of the Turkish text:

    "Everyone thought, conscience and religion, the right to freedom is the right religion or belief, freedom to change religion or belief alone or collectively, in public or private, teaching, practice, worship and observance of the freedom of Izhar contains."

    Quite frankly, when I look at Google Translate's translation from French, it makes me wonder why French into English human translators are still in work! On the other hand, when I compare this with Google Translate's effort from Turkish, it makes me think that there will still be work for at least another decade for human Turkish into English translators.
    Collapse


     

    Aniello Scognamiglio (X)  Identity Verified
    Germany
    Local time: 12:34
    English to German
    + ...
    Long live transcreation! Oct 7, 2009

    Nicole Schnell wrote:

    I am paid for writing, not typing. And for transcreation rather than "putting words into German".
    Long live translation!


    Transcreation? You hit the nail on the head, Nicole!
    I suspect we'll be seeing a lot more of transcreation in the years ahead. Why?
    Because 'translation' sounds like a commodity; *transcreation* sounds like a service.

    ... and to answer Henry's question: I do not use MT (yet).


     

    Laurent KRAULAND (X)  Identity Verified
    France
    Local time: 12:34
    French to German
    + ...
    Tempted, but to no avail Oct 7, 2009

    Aniello Scognamiglio wrote:

    Nicole Schnell wrote:

    I am paid for writing, not typing. And for transcreation rather than "putting words into German".
    Long live translation!


    Transcreation? You hit the nail on the head, Nicole!
    I suspect we'll be seeing a lot more of transcreation in the years ahead. Why?
    Because 'translation' sounds like a commodity; *transcreation* sounds like a service.

    ... and to answer Henry's question: I do not use MT (yet).





    Transcreation sounds like being the implementation of the skopos theory. I can agree with that. As per MT, I was tempted to use a professional MT system but I am still waiting for something which would look like an answer from the publishers I contacted regarding technical and linguistic questions. And sorry to say that, but there is apparently not much going on beyond selling the programs in question...



    [Edited at 2009-10-07 09:21 GMT]


     

    Jeff Allen  Identity Verified
    France
    Local time: 12:34
    Multiplelanguages
    + ...
    about MT Turkish>English Oct 16, 2009

    French>English v. Turkish>English
    Tim Drayton wrote:
    ....Quite frankly, when I look at Google Translate's translation from French, it makes me wonder why French into English human translators are still in work! On the other hand, when I compare this with Google Translate's effort from Turkish, it makes me think that there will still be work for at least another decade for human Turkish into English translators.


    Tim,

    An answer can be found in the following.
    FR>EN has been worked on for over 3 decades by Systran and also for 10-15 years by a few other commercial MT vendors.

    Yet, which vendor currently provides a Turkish>EN MT system?

    It requires a significant amount of funding (time, specialized human resources, equipment) to create a new language pair. 2-3 years of full-time work with a team of linguists, programmers, and other specialists. At a panel session of MT vendors in 1998, the figure of 500K USD for a new language was stated. It does not seem to have changed much in the past 10 years.
    Due to this kind of investment, the MT vendors choose new languages carefully based on estimated general market need. If a corporate user wants to help fund a new language that isn't top priority, then that wouldn't necessary get refused, if the necessary funding were available. Again, see the amount cited above. It's not just a simple small proof of concept with minimal funding for a new language pair.

    So, it's not surprising that the language pairs often available by commercial MT vendors are usually the same, and seldomly cover less-prevalent languages.

    And that is just to get the first commercialized version in place. It needs ongoing maintenance to improve it. That costs money.

    As for Google's Turkish version, the quality provided for all the language pairs is not necessarily what can be considered as marketable for commercial purposes. Think of it as just a industry research project making the system available for free. Univerities do this all the time. Yet, it's far from having the same serious investment as off-the-shelf products that get sold to general customers or corporate users.

    The Statistical based MT approach has helped in some ways in order to provide a faster ready-to-market system.

    Another factor is parallel corpora for system training and testing. We conducted several projects in parallel with the same language independent MT system.
    Based on existing bilingual corpora:
    EN ES system (speech to speech MT) took 48 hours to build demo.
    EN KR system (speech to speech MT) took 6 months to create and align content based on 400 speakers and build demo
    En Haitian Creole system (speech to speech MT) took 9+ months to create and align content based on 300+ speakers and build demo

    If 10-15 major corporate customers wanted to seriously invest in a new EN Turkish language pair, it would certainly get some attention.

    Chinese has obviously been a priority over the past couple of years due to the Olympics.


    Jeff


     

    Jeff Allen  Identity Verified
    France
    Local time: 12:34
    Multiplelanguages
    + ...
    Laurent, escalate your inquiries about MT Oct 16, 2009

    Laurent KRAULAND wrote:
    As per MT, I was tempted to use a professional MT system but I am still waiting for something which would look like an answer from the publishers I contacted regarding technical and linguistic questions. And sorry to say that, but there is apparently not much going on beyond selling the programs in question...


    Laurent,

    which MT vendor(s) is/are not answering your inquiries?
    See a few MT user groups (no such similar support groups specifically for MT here on ProZ)

    http://groups.yahoo.com/group/PROMT_users/

    http://groups.yahoo.com/group/SYSTRAN_users

    http://groups.yahoo.com/group/Reverso_users/

    The question of customer support response time has come up a few times. Some tips have been provided on how to ensure that the request to the MT vendor was actually submitted correctly and received (by an ID from MT vendors customer support system) and additional escalation info when there is no reply.

    Don't hesitate to escalate your inquiry if you are not getting a reply.

    Jeff


     

    B D Finch  Identity Verified
    France
    Local time: 12:34
    Member (2006)
    French to English
    + ...
    Is it ethical? Oct 17, 2009

    I am concerned about the ethics of using MT when presenting oneself to the Client as a human translator. I have recently put in an application to be on a panel of translators for a Regional Chambre de métiers and their tender document specified that use of MT was not acceptable. As they asked for detail of translation methodology, I thought it best to explain the difference between CAT and MT. Unfortunately, it sounds as though this may be becoming more blurred.

    BDF


     

    Russell Jones  Identity Verified
    United Kingdom
    Local time: 11:34
    Italian to English
    First attempt Oct 17, 2009

    Having been urged by Henry at the Ohrid conference to "give it a go", I am currently trying it for the first time (for many years) - only Google, nothing fancy.

    Frankly, I am amazed. Everything needs editing to a greater or lesser extent of course but there are some really excellent phrases that bear no relation to standard dictionary entries and that would probably never have occurred to me when working at normal pace.

    My only gripe so far is that there is no consiste
    ... See more
    Having been urged by Henry at the Ohrid conference to "give it a go", I am currently trying it for the first time (for many years) - only Google, nothing fancy.

    Frankly, I am amazed. Everything needs editing to a greater or lesser extent of course but there are some really excellent phrases that bear no relation to standard dictionary entries and that would probably never have occurred to me when working at normal pace.

    My only gripe so far is that there is no consistency of terminology; an excellent idea in one segment is replaced with something literal or banale in the next.

    It is a little early to judge speed but, being much more adept with the mouse than the keyboard (I'm using DNS as well) my first impressions are that it is a little faster and less physically demanding.
    Collapse


     
    Pages in topic:   < [1 2 3 4 5 6 7] >


    To report site rules violations or get help, contact a site moderator:


    You can also contact site staff by submitting a support request »

    '[MT] is most often used alongside [TM] as an adjunct to human translation'. Are you using it?

    Advanced search






    Wordfast Pro
    Translation Memory Software for Any Platform

    Exclusive discount for ProZ.com users! Save over 13% when purchasing Wordfast Pro through ProZ.com. Wordfast is the world's #1 provider of platform-independent Translation Memory software. Consistently ranked the most user-friendly and highest value

    More info »
    CafeTran Espresso
    You've never met a CAT tool this clever!

    Translate faster & easier, using a sophisticated CAT tool built by a translator / developer. Accept jobs from clients who use SDL Trados, MemoQ, Wordfast & major CAT tools. Download and start using CafeTran Espresso -- for free

    More info »



    Forums
    • All of ProZ.com
    • Term search
    • Jobs
    • Forums
    • Multiple search