Has using DeepL become the norm? How to calculate PEs?

Jo Macdonald wrote:

For a simple, widely-used term where the English abbreviation is already used in the Italian "Collaudo in fabbrica (FAT)" DeepL suggests "Factory Testing (FAT)".

This might look OK and be good enough for some, but even if you don't know the terminology, a quick look/check shows FAT is obviously Factory Acceptance Testing.

Justin Trumain wrote:
Full article: https://www.languagecrawler.com/2021/04/a-state-of-flux-difficult-and.html

dieuebersetzung wrote:

This is not about using an online dictionary vs. a hard copy or about using MT to help unscramble a sentence...
It's about having a 50 page document that would take 7 days to translate using human translation vs. no time with MT and then maybe a day's work for PE.

Hans Lenting wrote:

The thing is, that even a target segment that describes exactly the opposite of the source, can sound valid.

dieuebersetzung wrote:.
Where is this industry heading?
I mean MT and PE are certainly very useful for many situations but is it ok to use it without making I t clear?

Kristina Love wrote:

This discussion has been a real eye-opener for me as someone new to the profession.

I don't think there's any question that what the OP experienced, just as he described it (and I do not get the impression that he is mistaken or exaggerating), was very unprofessional, lazy, sloppy work at best and fraud at worst. If DeepL was utilized for the translation and only 10% of it was edited to make it different, there is simply no way in the world that a quality translation was produced. That is to say, it might turn out "OK" for some practical purposes, but that doesn't really matter. What the client is paying for (whether the translator charges a low rate or a high rate) is the translator's very best effort to produce the highest quality translation that they personally can possibly render.

However, I'm with the people who say that as long as the translation is of high quality (and I would add, that it is of the highest quality that the translator in question is capable of producing, giving sufficient time, attention, and care to their work), it shouldn't matter which tools they used along the way.

I think the OP has very valid reason to feel cheated in this instance. For me it's a shocking story, because it sounds like these were translators who were not just some fly-by-night "randos" but rather people he had trusted in the past, or at least who had proven years of experience as well as excellent marketing, and possibly other impressive credentials. This is not what I expected to hear, and I now realize I need to adjust my view of the people who appear to be the "ideal" translators and seem to have it all. I guess all that glitters is not gold.

Maybe these were well-intentioned people who mismanaged their time by mistake and took on too many jobs at once and found themselves in a very unfortunate predicament that sadly led to some corner-cutting and ethical compromises. We don't know what is going on with them - did they get greedy accept too much work, or were they experiencing dire financial problems? I don't want to judge them, but what they did, they shouldn't have done, and should never repeat it.

This thread has revealed that there isn't much agreement about how editing and MTPE rates, or translating with memory tools, or translating with the assistance of machine translation, or with CAT tools, should be priced.

Adieu and others already pointed this out, but I feel it really needs to be stressed: I think the OP has a few misconceptions of all the differences between these scenarios and what kind of work goes into each one. Is the OP a translator himself in addition to being/working for an agency?

It's so important to be knowledgeable about this and to have clear communication and mutual understanding between client and translator before the project starts. I recommend Corinne McKay's book "How to Succeed as a Freelance Translator."

A CAT tool, or translation environment tool, can be used with or without translation memory (not the same as machine translation), and with or without machine translation. And of course translations can be carried out without a CAT tool. In addition, as I believe a few others pointed out, Machine Translation Post-Editing is when the client gives the translator a project that has already been translated by a machine and then the translator "fixes it." If any kind of editing is done, it needs to be specified and agreed on: will the translator be editing a translation by looking at both source and target, or just the target? And will extensive revisions be necessary, or is it something that is practically ready and just needs to be proofread for minor mechanical errors?

Corinne McKay's book says that in general, the price charged for editing should be about a quarter to a third of what is charged for translating. I want to follow all of the advice in that book, but this one is a little difficult for me as I can see how "editing" is a broad term and may involve work that is almost or just as hard as translating.

I don't think there is anything wrong with using any and all tools. They can be used wisely, or they can be misused. Not wanting a translator to use a particular tool can somewhat be compared to insisting that they not use dictionaries or reference books, or insisting that they write it by hand instead of type. It may sound silly but it's the same principle. The tools are there to help create the best final product possible. With a difficult or lengthy project, I have noticed that using MT can even act as insurance against my own mistakes. If the machine suggests something I don't expect, it jumps out and catches my attention, and then I can focus in on determining whether it's the machine or my own understanding that's at fault. And yes, it DOES save time, but let's be clear: I'd probably spend a whole 7 days at least on a 50 page project even with MT. As a matter of fact, for any project, I'd use all the time that I had left until the deadline, regardless of what tools I was using, because the goal is to make it the absolute best it is within my power to make it. Every word of every sentence needs to be second- and triple- and quadruple-guessed, and more if time allows. The time that MT saves is time that can be re-invested back into refining the project and making it even better. At least 80% of the words need to be either reformulated or require time to research and confirm.

Some projects may be easier than that, and some translators may be faster and more efficient with this with more experience, but it doesn't change the basic fact that an honest and diligent translator with any kind of work ethic is going to spend a lot of time and effort on a translation, no matter what tools they do or don't use.

And it goes without saying this doesn't mean that translators don't have the right to rest, take care of themselves, have a personal life, or insist on a rate that they consider fair and worth their time and effort.

German Dutch Engineering Translation wrote:



Jenny Zonneveld https://www.proz.com/profile/1542616 posted this example at Facebook of MT gone wrong. First translation by DeepL, second one by Google Translate:




[Edited at 2021-07-28 07:16 GMT]

Yaotl Altan wrote:

Check the class GT also changes the gender of the (female) Portuguese writers.

German Dutch Engineering Translation wrote:



Recently there were lists at FB showing how GT links certain professions to specific genders.