posted 03 May 2001 14:29            

I have read Mr. Michael Benis article “Is Translation Memory Losing You Money?” in Transfree 29 with great interest. I am glad some one raised the issue. Whether the users of TM systems must give discount for full and fuzzy matches or not? It is an interesting issue, but how this question has come to the surface at the first place?

We translators are not the only folk using Computer Aided application to make our work more qualitative, consistent and productive. For example, many years before any TM system was available in the market, design engineers and architects were using CAD (Computer Aided Drawing). As early as 1990 almost 50% engineers and draftsmen were using CAD all over the world. Today, those using drawing boards are in minority. Let’s see how it effected on freelance design engineers and draftsmen and those firms which render their services for preparing workshop drawings etc.

Before I become full scale freelance translator, I had been working in industry as a mechanical engineer, heading a R&D Department of an engineering firm. We used to source out excess work to freelancers. The quality of the work was the only factor which determined the rate of payment, not the time or method or tools being used to accomplish the work. We readily paid better to those who were using CAD, because the quality of the drawings thus produced was much better. They were, at the first place good looking, neat and clean, without traces of erases and corrections etc.; the lines drawn were uniform, text and numbers beautifully placed. It was irrelevant whether a particular drawing was made by altering an old one from archive and database or produced completely new. After all we were paying for the skill, knowledge and the responsibility of the design engineer and of course for the end product (drawing) which should be serving its purpose.

Now coming back to our question of whether users of TM must give discount or not. The basic question we must ask should be whether TM gives a better quality of translation or poorer. If the answer is that the use of TM or CAT substantially improve the quality and consistency of a piece of translation, and thus gives some added values to it then agencies and clients must be prepared to pay it. If it is not so, and all what is achieved by the use of TM is speed and higher productivity, while the quality of the end product is slightly damaged then there must be a good discount.

In my opinion quality is the only measure and all other arguments like “After all, it's the TM system that "translates" the 100% translations and repetitions, just as it saves you research and maybe typing with its fuzzy matches, so why charge full price?” etc. are irrelevant.

Of course, “quality” is hard to measure and even harder to prove. So perhaps we are left with the issue of productivity and other benefit of the TM. And, as Michael Benis has said in his article, why it should be anybody’s concern, how one translates.

On the other hand the quality of the translation depends entirely upon the expertise of the translator and his skill to prepare and use his database as well as his experience and knowledge of the art of translation between a certain pair of the languages. So, it is the skill and expertise of the translator, which is paid, not the time or tool which is used for producing the translation.

If. “it costs us twice as much as Microsoft Office and puts the brakes on our productivity while we learn how to use it, and only then we may gain from increased consistency, no duplicated research time and an overall increase in productivity of perhaps 15-30%” then I see no justification on behalf of translation companies to demand any discount and depriving us from the profit on our investment. If any translator is offering discount, it is his(her) personal matter. After all we also have very different rates.

It is pity that the discount or no payment for full matches are not made by direct clients, who are not supposed to understand translation business, but by the translation agencies which are run by people who are themselves translators and who know very well how much work is involved when “100% match or repetition may need to be translated very differently in different contexts” as Michael Benis has correctly pointed out. And they are well aware of the fact that “it can easily take longer for an experienced translator to "fix up" a fuzzy match than to just type in a new translation.” So their demand for no payment for 100% matches and discount for fuzzy matches is not because of any wrong conception but because they want to maximize their profit on account of freelance translators.

Last year I was contacted by an agency for a job of 280 000 words long set of manuals, which they would be sending me at an interval of three to four weeks during the next eight months. They exclusively wanted to give the job to one acquiring or ready to buy Trados Workbench. They set a low rate, only 0.068 US$ per word, which I was ready to accept, but then they informed me that full matches will not be paid. My response was: “then (as what is a full match is a disputable issue), send me the text eliminating the repeated full matches, leaving only what you are ready to pay for.” Why should I use my horse to pull their cart if I am not paid to do so?

In short, I agree with you (Alex) 100 percent, when you say “If you have to spend time looking at or working on something you should get paid for it - no argument, no discount!” Your idea of having a “standard” is great, as we don’t even have any standard for pricing our services yet. As I have experienced, translating companies are paying from 45 US$ to 150 US$ per 1000 words.

Nevertheless, we surely require some standard. And we are surely big enough, if only enough of us will also follow the standard ourselves, then eventually we might succeed to get it endorsed and accepted also by the industry.