Friends Don’t Let Friends Use Trados or Other Translation Memory Tools

 

Patent translators are often very stubborn people. Sometime it works to their advantage, and sometime it works to their detriment.

I knew a translator in San Francisco who in the late eighties refused to use computers. He had been translating Japanese patents since the early seventies on his typewriter and he really liked the mechanical action of working on his typewriter. There is something about the mechanical beauty of writing on a typewriter that disappeared when humans stopped using typewriters after about 100 years and replaced them by computer keyboards by the nineteen eighties. This guy, who had his quota of translated pages that he was able to produce on his typewriter in the morning so that he could take the afternoon off (he was also refusing to work in the afternoon for a long time), eventually had to throw in the towel of course, because his clients wanted to have the translations in the form of computer files, not as typewritten pages. So eventually he did buy a computer, but refused to buy a modem – this was still in late eighties before there was Internet. He was sending diskettes by Federal Express, take it or leave it. Eventually, of course, he joined the rest of the e-mail literate crowd and started sending files from his AOL address. It must have saved him a bundle on Fedex charges. Some things just took him a little longer than most people – about 10 years longer. Incidentally, the only reason why the customers put up with the quirks of this translator was that he was also very good. Was he doing it mostly for his ego? Who knows.

I knew another translator who never learned touch typing. Instead, he developed his own method for typing with 2 fingers only. He was very fast, probably as fast as anybody who is touch typing. Plus the entertainment value of watching him pecking away on his keyboard extremely quickly with 2 fingers was simply indescribable. I could go on and on about translators who stubbornly refused to adopt new-fangled ways and resisted … almost until the bitter end – but usually not quite. For example, this translator refused to learn Microsoft Word for as long as possible. I was using WordPerfect 9 up until about 5 years ago and converting files to Microsoft Word, although it was a waste of time, of course, and the format was sometime messed up by the conversion. But WordPerfect was not really just a word processing program to some of us back in the day, it was more like a religion. I still have WordPerfect installed on my computers although I never use it anymore.

But to this day I refuse to use a translation memory such as Trados or whatever other translation memory tools there may be. There are many other translators, mostly people who have been translating for decades, who share my opinion. I don’t think that this is an irrational decision, unlike some of the decisions of stubborn translators described above. First of all, I would have to spend several hundred dollars first to buy the software and then I would have to learn it. It would probably take me a long time to learn it. I don’t want to do that. But the main reason why I refuse to learn a translation memory tool is that I don’t think that it would be very useful for my purposes. From what I understand about translation memory tools, they are very useful as they save a lot of time when one is translating chunks of repetitive texts with a lot of technical terms in them that are difficult to remember, for instance in instruction manuals which can be very technical and very repetitive, for example when new editions of the same software package are updated every couple of years or so.

But even though patents can be extremely repetitive, the same thing is repeated over and over like a Buddhist mantra in the claims, in the main description of the patent, and for good measure also in the description of the effects of the invention at the end of the patent application, the repetitive character of the text in one patent application is not applicable to other patents, which are written by other authors (patent agents) and may relate to other fields. Especially for somebody like me who works from several languages (Japanese, German, French and sometime also other languages into English), it would be expensive and really messy if I had to keep different versions of translation memory tools in different languages on my hard disk.

Some translation agencies these days require translators to use Trados or similar translation memory tool software. They can then create a database of technical terms in different languages which can be shared with other translators and used on future projects (but of course, the terms are then not owned by the translators who created the database, but by the agency). Oh, well, fortunately for me, I mostly work for patent law firms and none of them ever actually asked me so far about a translation memory tool.

In the interest of full disclosure, there is one translation memory tool that I do use. I use yellow post-it notes, the broad size, which I stick on the bottom of my monitor. On the left I write terms that for some reason are hard to remember in Japanese or German and on the right side goes the word in English. When I am done with the translation, the post-it note goes into garbage.

This translation memory works very well for this patent translator. And I don’t have to share my precious database of technical terms with anybody. I just keep them in my head. And when I die, another patent translator will have to create his or her own databases in his or her own head. The database that’s in my head will be gone forever. But the translations will still be there, on paper and on Internet. The other day I was Googling something and came across a translation of a medical study that was identified as mine. I took a look at it, and it looked fine, except that I could not remember at all that this was something that I translated.

Maybe I do need a translation memory tool, but not Trados. I should probably eat raw carrots or fruits that stimulate brain functions or something. I’ll have to research these memory tools at some point.

UPDATE I:

A translator living in Germany e-mailed me this link to an interesting discussion about a fantastic claim that one can translate 34,501 words with Trados in 10 hours. It should be noted that the translator who made this claim is a “certified Trados trainer” and thus presumably she can make money from people who buy the software and need to learn it. I found out about this discussion too late to participate, but I would be highly skeptical of her claim.  I sometime translate very long Japanese patents with extremely repetitive passages. I remember a long patent that described in 62,000 words something that could have been easily described in 4,000 words. So obviously, I use a lot of cutting and pasting for this kind of work. However, because I have to proofread both the source and the target language and look very carefully for small differences (such as “widget flange a’” instead of “widget flange a””), I can almost never translate more than about 5,000 words a day even with this type of highly repetitive translation. If I try to push over the limit of 5,000 words, I know that I will start making mistakes, which could cost me the customer.

I don’t think that the human brain can process 34,501 words of a translation in 10 hours (in two languages it would be close to 70,000 words, right)? Machines can do it easily, but machines don’t care about mistakes (or anything else for that matter).

I think the fantastic statement above is just another example of insidious commercial propaganda of the kind that we are unfortunately exposed to here in America on a daily basis.

UPDATE II

Here is another argument against translation memory tools. I don’t have to use them because I mostly work for direct clients, usually patent law firms, rather than translation agencies. I am pretty sure that patent law firms would be kind of leery about translators who actually use translation memory tools to speed up the process. But many agencies require freelance translators to use a particular TMS, usually Trados. Once you use it, the agency (or even a direct client) may  create a sliding scale for rates depending on the number of “matches”, or repetitions of words and sentences, which means that you get paid less in the end if there are many repetitions in the text.

A few months ago, I translated two very similar Japanese patents for a very small agency. I have known the guy who runs it, who is a translator himself, for something like 20 years. I sent him an invoice for three thousand dollars. He e-mailed me back, saying that based on “matches” in the text of both patents, he would only pay me two thousand three hundred dollars. Well, some passages, such as prior art, were very similar, almost identical. But the final proofreading of these similar passages takes up so much time, precisely because they are so similar, that I have to charge full rate. In the end he did pay what was in my invoice, but only after I fired off a slew of very angry messages. And our relationship of some 20 years has been seriously compromised.

I think that this particular translator is used to this type of reasoning because that is what agencies and clients that he works for are used to as well, since he is using Trados or something like that. Translation memory software thus may exert a downward pressure on the rates of all translators, even those who are in a position to refuse to use them.

So after all, it might not be such a good idea to use Trados if you want to get paid a decent rate.