A discussion with Marc Prior


Dear Colleagues,

Many thanks to Antonio Oliver Gonzalez for the “virtual” invitation to Spain and to discuss software for translators with you.

I am of English origin but have lived in Germany for 20 years. I have been a professional translator for 24 years now, and for the last 18 years I have been a freelance. I translate from German, Italian and Dutch into English.

Around 13 years ago, I began using the Linux operating system, and I have built up resources for translators who use Linux at www.linuxfortranslators.org.

Since 2003, I have also been the project co-ordinator of the OmegaT project. OmegaT is the most popular free/open-source CAT (computer-aided translation) software. I have heard that OmegaT is also taught on courses at your university.

Now that I am here, please take advantage of this! My knowledge and experience are at your disposal. Don’t worry about making mistakes in your English – I don’t speak any Spanish, after all.



Is it possible to be a freelance translator using exclusively open CAT tools? I am aware that nowadays a lot of translation agencies prefer to send Trados projects to their translators. To what extent is an open CAT tool like Omega T compatible with the other expensive tools that a translator may be requested to use?


I don’t use any CAT tools besides OmegaT, and I make a reasonable living, so I suppose it must be possible! I am not the only one using OmegaT, either. There are even people who don’t use Microsoft Office, or who don’t use Windows.

The question though is not whether this is possible, but whether it is beneficial or advisable.

Compatibility with other CAT tools and with Trados in particular is not a black & white issue. There is some compatibility, but it is not perfect. (Even different versions of Trados are not completely compatible with each other.)

For example, OmegaT uses the TMX format for translation memories, and Trados is able to read or export translation memories in the TMX format. So if a customer has translation memories from previous jobs in TMX format and it sends you those TMX files, you will be able to use them in order to get fuzzy matches or perform searches. Equally, if you do a job for a customer, you can supply the TMX file from the OmegaT project to the customer along with the translation, and the customer will be able to send that file to other translators in future for use as reference.

OmegaT can also handle bilingual files prepared in Trados in the TTX format (provided attention is paid to certain points), and can handle Trados’ own SDLXLIFF format. But there are some limitations, and coping with them means that there are more things to consider.

You can find more information here: www.omegat.org/en/howtos/compatibility.html.

Reading and understanding documentation like this becomes much more important if you are doing something differently to everyone else.

It is very important that you understand what exactly your customer wants you to do, and whether you are able to do it. It is not enough to know that your customer “uses Trados”. Your customer may wish to supply you with a Trados TTX file for you to translate, for example. In that case, you can offer to translate a few sentences in OmegaT and send the TTX file back, to show that you are able to handle the process. Besides acquiring the technical skill to do this, you may also need to hone your skills of persuasion, since many customers will simply not consider you if you don’t have Trados.

When you consider these issues, it’s easy to understand why people simply pay the €850 for Trados. (Which, incidentally, I do not think is a lot of money. I bought Trados myself, in 1997, when it had far fewer features, for 4,000 Deutschmarks (around €2,000).)

I suspect that most OmegaT users do not require strong technical integration with their customers. Most of my customers for example just require the translation. With a few of them, I exchange TMX files. But I have never handled a TTX or SDLXLIFF file for a customer, even though this is possible.

I feel that the better customers are more interested in the quality of my translation than the tool I use. But not everyone would agree, and if you are new to the profession, you may not be able to be too choosy about your customers.

Incidentally, the issue of compatibility between OmegaT and Trados is similar to that between other tools, such as Déjà Vu or Wordfast, and Trados. It is not particularly an open-source issue.


I don’t think I exaggerate when I say most of us have been struggling with the technical part of this course, not many of us has much of an IT background. What would you consider necessary for a translator to become “self-sufficient” (I would like to be self-employed some day) in creating resources for him/herself? What would you start with?

In my opinion, OmegaT is not difficult to use if you have a good basic knowledge of IT. For example, knowing how to use your operating system (Windows, Linux, Mac, whichever), basic “housekeeping” tasks, like being able to find/manage/organize/move files/folders, edit files, install programs, etc. However, translators often lack even these basic skills, and then learning a tool, like a CAT tool, becomes a problem.

I learnt to use a computer in the late 1980s, before Windows appeared. In those days, you had to learn how to use a computer properly, and to understand exactly what you were doing. Now, with Windows and the graphic user interface, you can often “guess” what to do. Unfortunately, it is not always possible to “guess” everything correctly, and in my opinion, people need to learn again to read instructions and follow them. (Which is good for translators of course, because someone has to translate those instructions.) IT knowledge is not the most important thing about being a translator, but I think it is going to remain a vital secondary skill for the foreseeable future.


My first question is similar to that of Christian but focused on translation agencies: are there real statistics about the spread of OmegaT in the translation market? That is, is it specially designed for freelancers or is it also widely used by translation agencies? If so, is it possible to set up an agency that works only with CAT tools? After finishing my studies at university 7 years ago, I did an internship in a translation agency and I was introduced to Transit and Trados but nobody mentioned the possibility of using an open CAT tool. In fact, I think they didn’t even know about it.

The number of users we have is something that interests us as well. Most CAT-tool vendors can at least say how many copies of their product they have sold, but since anyone can download OmegaT free of charge, we can’t do that. There have been hundreds of thousands of downloads, but that is fairly meaningless.

So, we only have “circumstantial evidence”. For example:

You can compare the number of members of different user groups, such as: tech.groups.yahoo.com/group/omegat (1684)
tech.groups.yahoo.com/group/dejavu-l (2391)
tech.groups.yahoo.com/group/wordfast (6768)

This is not necessarily a reliable guide, since there may be more than one user group for some products.

You can also compare the number of people declaring usage in their ProZ profiles. For example:
Go to www.proz.com/translator-directory/
Then Select English as the source language, Spanish as the target language, OmegaT as the Translator productivity software, and view the results. Then repeat the search for a different CAT tool, and compare.

You must also consider that people who advertise that they have OmegaT may not actually use it, especially if they list other CAT tools. Few people will pay €850 for Trados, or even more than that for Transit or Déjà Vu, just to say that they have it. But some people almost certainly advertise that they have OmegaT, because it doesn’t cost anything to do so and so they have nothing to lose.

Based on indicators like these, I estimate that OmegaT has about 3% of the CAT-tool market. Since we estimate that Trados has about 70% of the market, that would mean that we have around 10% of the rest, which we think is quite good considering that we are not a commercial company.

OmegaT is specifically designed for the people who produced it, and they are freelancers. A few translation companies may be using it, but not many. But that doesn’t mean that you can’t use it and work for translation companies. It depends upon whether they have an integrated workflow, and if so, what it looks like.

OmegaT was fairly new seven years ago; it didn’t really appear until 2003. So it’s not surprising that the translation company you worked for hadn’t heard of it at that time.

As I mentioned in a previous post, it is quite possible to work as a freelancer using OmegaT, and it is also possible to work as a translation company using OmegaT. The question is just whether that is the best thing to do. (And there is no simple answer to that question.)


In the last few weeks, we have been working in open source/free resources for translators, how to create them or to find the appropriate ones for us. We have worked with OmegaT, I guess my question is, how did you get involved in developing a project like this? And what kind of training you had prior to developing the program and acquired in the process of creation of the programme?

I have never been involved in developing OmegaT, i.e. writing the code. (I hope your tutor does not now think that he has invited the wrong person…) The original OmegaT program was written by Keith Godfrey, about ten years ago, and when I joined “the project”, there were just the two of us, Keith (who wrote the software code) and me (who did everything else). Since then it has been extended and improved by many other developers, about 24 according to this page: www.omegat.org/en/who_we_are.html.

We do not have top-down management like a commercial company. I have several roles, but one role is to try to keep track of what everyone is doing, and so I have the title “project co-ordinator”, and not “project manager”.

I did not become interested in OmegaT because it was open source. I was (and still am) quite prepared to pay money for good software, and I do not attach importance to having access to the code myself, since I would not be able to do anything with it.

I became interested in OmegaT simply because in 1999, I began using Linux. I wanted to use a CAT tool, but at the time there was no CAT tool for Linux. So when OmegaT appeared, which ran on Linux as well as Windows, I was interested, and I got involved and advised Keith on how to improve it.



I’m also curious about the first steps when becoming a freelance translator. What would you recommend? Sending loads of CVs to translation agencies or maybe targeting those companies that might need your services?

The first thing, more important than anything else, is to realize that becoming a freelance translator means setting up a business. You must become a businessperson. Most of your customers will be other businesses, so for the most part it will be a business-to-business service (although there are exceptions).

Then you must decide what service you are offering. Translations, of course, but from/to what languages? Do not say, for example, “I will offer French to Spanish translation because French was one of the subjects I studied”. You must think like a businessperson. Is your understanding of French good enough to offer French to Spanish translation as a business service? If not, you don’t even need to mention it to customers. Don’t mention your hobbies of horse-riding and clarinet-playing, either, or what school you attended, or anything else that isn’t relevant to selling your services. Businesspeople don’t do that. (At least, not in Germany or Britain, but of course there are cultural differences.)

If something is relevant, don’t leave it to the customer to guess why. Don’t say: “2005-2009: I worked for my uncle’s construction business in the summer months.” Say: “I specialize in translations for the construction sector, and I have gathered considerable hands-on experience spread over many years in this sector.”

You must of course decide what specialist subjects you are going to offer, and it can be difficult to get the balance right (too many/too few or broad/narrow).

Then there is the decision to support certain software formats, perhaps to offer to translate websites, to be able to integrate with businesses (usually translation companies, but also some end customers) who use a particularly CAT tool, and so on.

Whilst you are considering what services to offer, you need to consider who is going to buy them. What kind of business, what kind of sector, located in which country or countries, etc.

That of course will influence how you are going to approach these customers to sell your services to them. Yes, sell your services, not “apply for work”, because you are a businessperson, not someone applying for a job.

When you have done all that, my belief is that you will decide that a CV is not the best way to sell your services. A CV says: “This is me. Hire me!”. What you need is something that says: “This is the service I offer. If it is what your company needs, please contact me”.

Having said all that, from my own experience, the best opportunities are passed from one translator to another. Because it is very difficult for customers to judge how good a translator is, they often rely on the opinion of someone who can judge – another translator. So word of mouth is very important, getting to know other translators, networking.

You may still decide that you are going to begin by approaching translation companies, and that you need to send them something (and the reality is that translation companies usually expect a CV). But it is a good idea, in my opinion, not to lose sight of what you are trying to do, which is to sell your services, not “apply for a job”.

Another interesting answer to a very similar question:

First, if possible, I advise people not to become freelance translators if they only have an academic background (school > university > freelance translator). It is better to have worked first, whether as a translator or in some other job/profession. As I mentioned in a previous post, being a freelance translator means running a business, and some life experience is good, possibly essential for this.

Second, as I think I have also already mentioned, networking with other translators is very important. How you do this is up to you, but joining a professional association is one way.




“I’m also interested in how should translation services be charged? (…) but how can we make sure that we are being fairly paid and not selling our work at a loss?”

Firstly, you must start thinking like a businessperson, and do some calculations.

This is not necessarily easy. You can estimate that a job will take you four hours, but not realize that it is going to take you six hours until you are half-way through it, when you have already committed yourself to a price.

But with experience, you will become better and better at estimating an appropriate rate for your work, and you will gather more and more information on what rate a customer is likely to accept.


As for me, I have a degree in Philosophy and I am currently on a PhD program. I have some experience translating articles, papers, and projects for my department at Uni. At the moment I am working on the translation of a correspondence between two german (jew) philosophers (L. Strauss and J. Klein). As you can imagine, being philosophy my field of study – a rare specialisation in specialized translation – I find it very hard to come along resources for this kind of half-literary half-specialized translation.

I was already told at an interview that literary translation “is for rich people” – because it pays so bad. I’m sure the same applies for “philosophical” translation. And I am aware that the stuff we discuss here most of the time does not apply to this kind of translation. But I wonder – in how far? Have you ever translated anything of the sort? Do you know of any people who do, and if so, do you know what online resources they use?

I don’t remember translating anything that could be described as “literature”, though in 23 years of translating, as you can imagine, I have translated some strange things. I have translated speeches held by famous politicians (but not often). I even translated a love letter once. I have translated whole books, but only on technical or scientific topics.

I sometimes translate marketing-related material, which requires creative skills. Yesterday I translated an advertising slogan, just four words. It would be silly to charge for this by the word, and I negotiated a price of €75. That includes discussing the possibilities and implications with the customer. Some translators enjoy doing this, some like doing it just for a change from routine work, and some hate it. I suspect that those who hate it do so because it is time-consuming “per word” and not therefore lucrative if they don’t negotiate effectively.

As you see, I keep bringing the subject back to money. Although I love languages (I have learnt ten, including my native English), and I can’t imagine doing a job that I don’t enjoy, at the same time I don’t think that I would be prepared to accept poverty just in order to be a translator. It is a job that I enjoy doing, but it is still a job. The well-paid translation work is in commercial/industrial sectors where there is money to pay for it.

I suspect that just as many novelists struggle financially in the hope of having their work read by millions, the same is true of many literary translators. They have more in common with authors than with technical translators like me. In fact, literary translators are quite likely to be authors themselves.

So, to answer your question: I am the wrong person to ask. 🙁


I know my question is not strictly related to the topics proposed by Antoni Oliver, but having this opportunity to meet someone who has witnessed the evolution of the translation business over many years, I´d like to know your opinion about the impact of electronic books on the future of translation, both of literary and technical works.

Books have never been so cheap and easy to copy, and, unfortunately, we´re seeing that many people are unwilling to pay authors and publishers for their work, let alone translators. (In fact, I have the feeling that most readers don´t care at all about the quality of the translations they read).

As translation is such a time-consuming task and, unlike writing, one which is not likely to bring you fame, I assume not many people would do it for free if there is no demand for high quality translations. And the less speakers your language has, the worse things could go. Where is this branch of business heading? Do you think there will still be a market for (good) literary/technical translations in, say, 20 years? And if – as some propose – readers and authors keep in direct touch without the middleman, where does that let translators?


Firstly, I must say that I’m not really familiar with the literary translation market. My work is strictly technical, and mostly industry-oriented.

It is however generally known that most literary translators barely earn enough to live on, even those translating into English or German, which are relatively healthy markets.

I can only speculate on the impact of e-books. They might even increase the number of books being read and bought. However, it is a sad fact of life that when something becomes intangible, like software, a computer file containing a novel, or a sound file containing a pop song, it becomes devalued, and people are less willing to pay for it than for the tangible media of a book or CD.

It is also true that less value is being placed on good writing. It is quite common for example for a British newspaper, even a respectable broadsheet like the Daily Telegraph, to contain numerous simple errors, like the confusion of “its” and “it’s”, a mistake that a 13-year-old grammar-school pupil is not expected to make. Among the wider public, grammatically correct writing, never mind good writing, is becoming rarer and rarer, at least in English.

However, I see this firmly in the context of business. It is up to writers to “market” the value of good writing, just as the providers of other services and products have to convince their customers of their value. It isn’t enough to say “I am a qualified translator with degree X from university Y, so my translations must be good.” The customer must be made to understand the benefit of this to him or her. The best way of doing this is perhaps simply to discuss the text with the customer. When they are asked to explain the finer points of their own texts, customers and authors are often astonished at how much thought translators give to them.


I have one question about the relation between OmegaT and Okapi Tools. I think this is a very good combination of tools. Some weak features of OmegaT (as the possibility of using very big translation memories) are well implemente in Okapi (with the PensieveTM, for example). On the other hand, Okapi Tools does not implement any translation editor because OmegaT is a good choice.

I want to know if OmegaT and Okapi Tools are separated projects or if there is any contact between them.

OmegaT and Okapi Tools are separate projects, and both would exist if the other did not exist. There is however a high level of co-operation. A visible example of this is the Okapi Filters plug-in, which enables OmegaT to handle several file formats (notably Trados TTX) that it cannot handle natively. It is possible to do this by conversion routines within the Okapi Tools, and these may also provide more functions and better control, but the plug-in is very user-friendly for OmegaT users.

The Okapi Tools provide a large number of functions and features that supplement and extend those of OmegaT (and probably other CAT tools as well). However, I personally find the Okapi Tools difficult to use. This is partly because the documentation is geared to developers and experienced IT users, and not to ordinary translators.

Regarding very big translation memories, OmegaT usually handles these quite well natively. For example, look at www.translationtribulations.com/2012/01/recent-experience-when-tutoring-new.html. There, after testing, Kevin Lossner reported that “In terms of overall performance, the best results were obtained with OmegaT and “Trados Classic” (2007). Searching a huge TM gave results in a flash”.


As an experienced professional in translation I would like to know, how do you assure the quality of your translations? When you work directly with a client, and nobody can revise your work after you(corrector, reviewer). Do you have any specific methods you have developed to do your own “quality control”? Could you give us some useful tips?

There is no perfect “quality assurance procedure” that guarantees that a translation will be perfect. I think that everyone has to find their own method.

My procedure, regardless of whether I am working for an end client or a translation company, is this:

1. First draft, translation in OmegaT (I switch spellchecking off for this, because I find it distracting)
2. Second draft, also in OmegaT, comparing my translation segment-by-segment with the original (CAT tools are very good for this). I switch spellchecking back on for this
3. Print out, read through on paper, mark any changes on the hard copy, then enter them in OmegaT
4. Repeat (2.)
5. Spellcheck the translation in a word processor

There is no consensus. Some people insist on reading through the whole source text before starting to translate, and making a note of terminology issues as they do so. Some say that you *must* read through your translation on paper, whilst others have an entirely paperless procedure. And so on.

Also, although checking by a second person is always a good thing (called “a second pair of eyes” in the English-speaking world, or “das Vieraugenprinzip” in German-speaking countries), there is no guarantee that all errors will be found this way. Indeed, in my opinion a translation by a skilled, experienced translator, but not checked by a colleague, is likely to be better than one produced and checked by two different translators who are not as experienced. By no means all translation companies check translations from freelancers; of those that do, many only perform rudimentary checks. If the text is highly specialized, in particular, I think it is comparatively rare for a translation company to commission an expert translator in that particular field to do the translation and another expert translator in the field to do the checking.

Here are some tips:

1. Only translate what you actually understand, and if you don’t understand the text, expect to have to do a lot of research until you do. I spend much more time researching than translating, even though I only work in certain fields. Most commercial translation is specialized in some way, and one of the main reasons for bad translation is that the translator is not conversant with the specialism.
2. Put time before drafts if possible. Ideally, don’t do a job entirely in one day; a night’s interval between two of the drafts is good.
You may find that you have trouble code-switching. In other words, you are always thinking either in the source language, or in the target language, and you cannot switch repeatedly. This is true for me. When I produce my first draft, I am thinking in my source language (usually German). Often I know *exactly* what is meant, but try as I might, I can’t find the best target (English in my case) rendering, so I put something down that is “pidgen English”, or a literal translation. When I have finished the first draft, I go away, have a coffee, take the dog for a walk, have a night’s sleep, etc. Then, during the second draft, I am thinking in the target language (in my case English) the whole time and a phrase that I spent 15 minutes during the first draft trying to think of occurs to me immediately. This may not be true for you; but it is certainly true for me.
3. Try doing a translation and checking it, leaving it for a few days, checking it again, and repeating the process several times. Do this with a few different texts. Eventually you will realize that there is a certain number of checks beyond which you are unlikely to find any more mistakes or make any more real improvements to the text. (Although if you then give the text to a colleague to check, you may still be shocked by the results.) This is how I arrived at my procedure (see above): for many years, I only translated, checked against the original, read through (and then spellchecked), but I eventually decided that another round of checking (stage 4.) was appropriate.
4. Don’t believe anyone who claims to have “found all the mistakes” (in their own text or yours). Perhaps they have, but how can they know?
5. Experiment. For example, try changing the font on your screen (or on paper) from time to time. Try reading your translations out loud to yourself.
6. In commercial translation, a text has to be “fit for purpose”. It has to be good enough to do the job. Always keep this in mind. It is extremely difficult to produce a “perfect” translation, but if you remain conscious of what the text is going to be used for, you have a much better chance of producing an adequate text. For example, you will often find phrases that are ambiguous, and interpreting them correctly will depend to a large extent on your understanding of what the text is going to be used for.

Non-professional translators who use OmegaT: yes, occasionally we hear about them on the user list. It also depends I suppose what you mean by “non-professional”. For example, volunteer translators, or people translating just for fun or in order to learn a language, etc. I went to Hungarian evening classes for six years, and when our teacher gave us Hungarian translations to do for homework I did them in OmegaT. 🙂

Based upon what I have seen of so far, OmegaT appears to be good enough/sufficient for a professional to use as their only CAT tool; do you agree?


OK. First distinction: professional vs. commercial. Is OmegaT good enough for *professional* use? Or, putting it differently: is it possible to translate to a professional standard using OmegaT?

Certainly. There are however some limitations that need to be considered. For example, the lack of a split/merge segments function can constrain the translator at times, and make it difficult to restructure the text, for example when the translator would prefer to reverse the order of two sentences. On a more general level, it could be argued that presenting one segment at a time encourages the translator to think of a text as a collection of unrelated segments rather than as a cohesive whole. (This second point applies to all translation memory tools, not just OmegaT). However, it could also be said that a true professional should be conscious of these issues and not allow themselves to be constrained by the tool’s limitations. There are workarounds that can be used to overcome the tool’s limitations. Some of these workarounds are quite clumsy, but they don’t affect the final translation (i.e. they enable the translator to translate precisely as he or she wished to translate and would have translated if not using a CAT tool).


Some comments on other questions about the capability of OmegaT to handle very big translation memories

“Regarding very big translation memories, OmegaT usually handles these quite well natively. For example, look at www.translationtribulations.com/2012/01/recent-experience-when-tutoring-new.html. There, after testing, Kevin Lossner reported that “In terms of overall performance, the best results were obtained with OmegaT and “Trados Classic” (2007). Searching a huge TM gave results in a flash”.”

This is a very good think to know about OmegaT. I had never heard of MemoQ and it also looks interesting. I found useful Kevin Lossner’s comparison between the different tools and how they display concordance results.

I thank both of you for writing about this.


I am not professional translator, I am a teacher, but I am doing this course to learn to use some CAT tools, just in case I have to do some translation (not professionally). I really like OmegaT because it’s free and I agree with you that is not difficult to use if you don’t have much ID experience. The only problem that we have now is with Google Translate, because you have to pay now to use it with OmegaT. Well, Antoni explained us how to do it, we have to follow the instructions in the following page:
I know that you cannot trust automatic translation but sometimes it helps a lot! For me, it is also useful because I translate some archives from German into Spanish/Catalan and Apertium doesn’t offer these languages.
In fact, it is not so expensive as buying a payable CAT tool. Would you recommend us to use it?

Yes, as you say, the Google Translate situation is explained at www.omegat.org/en/howtos/google_translate.html. In fact I think this page is currently being translated into Spanish.

I don’t use Google Translate myself and I don’t think that using it will become universal among translators, but I accept that it can be useful. I think of the benefit more as having a very large translation memory to refer to. That “translation memory” may contain terms or phrases that the translator might not otherwise have thought of, for example.

I wouldn’t recommend using Google Translate, or not using it. If you do use it though, don’t rely upon it; just use it as a tool. Check and double-check everything. It is like a bad dictionary in a way, not reliable but useful if used carefully.

Although you now have to pay to use the API, i.e. Google Translate from within OmegaT (you can still paste text into the Google Translate website free of charge), if you use it professionally the cost is in fact very low. So if you think it’s worth using GT, it’s probably worth paying the small fee.