THE HOW-TO LIBRARY

Tool Kit * Are We Not Profitable?

Jos Bailleul, one of the developers behind the then very promising TEnT (Translation Environment Tool) TransSuite 2000 published the professional farewell letter:

(…) Knowing that almost 70% of our TRANS Suites that are currently running around the globe are illegal, I can say that a lot of you helped us to reach the point where we had to close the company because we had lack of funding. For some clients we had to give up to 15 PIN codes for ONE license in a year, which does say enough about the number of illegal licenses they were running.

It sort of shook me up back then because it displayed in such stark terms that our poor reputation in the software industry for not being particularly honest about how we use and pay for software is actually a fact. After reading his letter, it wasn't surprising to see that the software vendors that cater to the translation industry were among the first to roll out the dongle attack in full force—which of course now has largely given way to other forms of secure licensing.

So, there it is. A good portion of the many thousand translators around the world are simply using software for free (assuming that things haven't changed much since the demise of TransSuite). But is this the only reason why the market for translation tools is so relatively small?

Here's what industry analyst Don DePalma from Common Sense Advisory says:

The entire market for translation automation and localization tools, including translation management and machine translation, was somewhere in the neighborhood of US$100 million for 2006. When I followed the application development tool market back in the 1990s, you could find any number of single tools doing $10, $25, and even $100 million per quarter—that's for just one product, not for an entire industry.

And note that these numbers include many more (and more expensive) tools-such as localization, machine translation, and management tools-that are not part of the two dozen TEnTs that we tend to think of.

At a conference a few weeks ago, I had a very interesting talk with a gentleman from Temis, the company that bought Xelda, the terminology suite developed by Xerox's French subsidiary. Those who have never seen this tool in action would probably be just as amazed about its power and accuracy as I was when I first saw it. Among other things, it extracts matching term pairs from a pair of documents or a bilingual file better than any other of its competitors. And the reason for this is not surprising. In contrast to its competitors, it uses actual linguistic data to match up terms rather than mathematical formulas. Of course, this also means that the number of supported languages is limited, but be that as it may. I am not writing this to convince you to purchase the tool, because that is actually rather difficult. While Temis does still officially sell the tool, we (the translation industry) are simply not interesting enough for Temis to invest in its sale. In other words, we are not willing enough to pay for products to convince vendors like Temis to proactively market their tools to us, let alone continue to develop tools for our needs. Instead, the functionality has now been repackaged and integrated into Temis' powerful search engine tools that it is selling to the information industry.

I'm certainly not writing this to blame Temis—why in the world would they try to sell a product to a market that is simply not willing to pay? But something is wrong when there actually are products out there that are better than any other existing tools, but they are not being marketed or developed. Do I have an answer to this conundrum? Well, not really, but I know for sure that our reputation is not helpful. Neither is our general unwillingness to invest in the tools that are—or could be—out there, choosing instead to carry out our work with the same old manual and time-consuming routines.

Other possibilities out of the quagmire might also include new business models from the vendors. Lingotek and across—and, of course, the open source tools—are already offering different approaches with some free versions; Elanex TSE is going a new and different route with a barter model (you pay some translation memory data and get some in return); and Beetext is now offering its tools in an SaaS (Software as a Service) model with a monthly license fee. So, we don't necessarily have to continue with the old buy-and-upgrade model, but if we want changes and access to tools that we otherwise may not get access to (even illegally, because they may never be developed), then we need to be flexible (and honest) as well.

After I first published this article, I asked for responses. Here are some excerpts, first from a couple of tool vendors themselves:

Just a quick note to the point you raised on profitability: I also believe that one of the reasons Product X got so popular is the fact that it has been around in a relatively easy-to-access cracked version. I never saw anyone using a cracked version of any other tool. I talked to quite a few companies recently about this, and despite the fact that they are using a licensed version of Product X, they all have the crack to it somewhere -- just to use it as a temporary solution if they run out of licenses. Having a working crack gives a feeling of security, and on the other hand, heavy protection and dongles and stuff really do bad to those users who pay, not those who don't. However heavy your protection may be, a cracker who wants to crack a tool will crack a tool. Probably the others are not available in crack because there is not enough interest in this market.

This was István Lengyel from MemoQ (I edited out his specific reference to the product name). As an additional note: unfortunately there are cracks (little illegal software programs that crack the copy protection of the tools in question) for a number of tools out there, not just for the one.

Here is what Rodolfo Raya from Heartsome had to say:

I have a brief experience working with translators, but what I saw in the last five years makes me think that the numbers given by Jos Bailleul were not exaggerated. Some people purchase one license and try to install it on three or more computers.

Let me play with the math. If we have 10% of pirate users and each of them is responsible for five illegal copies, then it means that for every 1000 licenses sold there are 500 illegal copies. In other words, if 10% of people give out five copies, one-third of the installed base is pirated. I want to make it clear: only a minority of translators cheats, but the impact of that minority on the software development industry is terrible. Piracy affects software sales and has some side effects that people usually don't consider.

I'm currently working in China and what I see here is amazing. You can buy any program on the streets for one dollar. Translators here use any CAT tool because the price of any brand is the same for them. Pirated software is better than open source alternatives because it is readily available and there is no need to wait for a download.

Now, I happen to know that both István and Rodolfo feel very passionate about their work (actually, Rodolfo ends his note by saying that "writing software for language professionals is a gratifying experience and the existence of a few bad guys is nothing compared with the vast number of excellent people I met in this industry"), but, boy, this is not a good situation and does not bode well for future development.

There were lots of other comments as well. They ranged from the need to create open source models to a more realistic price policy on the side of TEnT vendors to better training opportunities. And these are all good and valid points. But there was one statement that summarized my own opinion very well:

Perhaps translators need to take a more business-like approach to their work? You can't economize on (or rely on a pirate copy of) something that ensures your business's survival!

Recommend this article: ||||