Things have gotten a little quieter around Ukrainian translation application powerhouse AIT in the last year or so. Awhile back it seemed that there was a new product and/or release every couple of weeks, whereas now it is usually a few months before a new announcement makes it to our attention. But maybe even these inventive guys need a rest occasionally. After all, they have come up with numerous solutions for counting (various versions of AnyCount, ClipCount, or CATCount, a tool that converts complicated TEnT count numbers into accounting-friendly formats), terminology (AnyLexic, AcroLexic, and WinLexic), time management (ExactSpent), project management/accounting (Projetex for the LSP and Translation Office 3000 for the freelancer), and most recently a translation memory solution (AnyMem). You'll notice I don't use the term "TEnT" here for AnyMem because it is not a complete translation environment but truly a translation memory-centric tool. I'm looking forward to the day when they release the product that ties (almost) everything together into one bundle, something I know they are working on (which also may be the explanation for the relative quiet recently).

I often get emails from readers telling me about this or that word count tool they just found and asking me to mention it. I usually don't. There is such a plethora of translator-specific word count tools out there, it seems to me that mentioning all or even many of them just adds to the confusion. And confusion there is.

So, do we need word count tools? Yes, we do. I realize that most Office applications and all TEnTs count words, but since they all do it differently, it really muddles our communication with clients who typically rely on, yes, you guessed it, Microsoft Word. So, for mere word counts (excluding repetition and TM match counts), a tool that can emulate what MS Word does while at the same time not committing its sins (like not being able to count stuff that can't be opened in Word or foregoing certain parts of files), fulfills the first set of criteria. Now, what about repetition counts or counts of matches from the translation memory? This is what we have TEnTs for. If you and your client use and trust the same TEnT, you should probably just use what those counts say. If you and your client don't agree on a tool or if only you have a TEnT, then I also would suggest making a count in the TEnT. But rather than taking the actual numbers, I would just take the percentage of repetitions and matches and apply that to the "correct" count of your Word-emulating counting software.

And, yes, it's a couple of extra steps, but we are only talking about a few minutes here (if that) and those minutes might go a long way toward avoiding conflict later on.

So, what else should be a criterion in selecting a counting tool? The number of formats it supports. MS Office, HTML, XML, text, and PDF files have become a given as supported formats in these tools (though every PDF count should be taken with a grain of salt since text might actually be included in non-countable graphics!), and more exotic formats such as compiled help, OpenOffice, and FrameMaker are also supported by a number of tools. However, after that it becomes really exotic, and that's where the wheat separates itself from the chaff. The latest version of AIT's AnyCount supports some surprising formats, including MS Publisher (I'm not sure I want to know how many words these files have since I don't want to translate them in the first place -- real pain!) and, more interestingly, graphic formats. To be able to do this it uses an internal OCR engine (which makes the download 10 times as large as it used to be) and the results are . . . mixed.

I ran a number of tests. Some counts were very accurate, especially on graphics with quite a bit of text and a normal text flow. With some, such as my company's logo (with "International Writers' Group" in small caps), it went ballistic: it actually gave me more words than there were characters!

So, overall this is an interesting new feature -- I have spent too much time manually counting words on graphics for projects to not appreciate it -- but you really want to make sure you know which kind of text/graphics are counted well before you put all your eggs in the AnyCount basket.

One other caveat is that the image counting feature is only available in the much more expensive Enterprise version -- but I'll talk to the guys from AIT and ask whether they might offer us a special in the next newsletter.

A number of years ago, I ran this article in my newsletter:

One of the more time-consuming and unnecessary problems in communicating with clients is disagreements on linguistic renderings for which there are several "correct" solutions. Typically these disagreements come once the project is finished and often require a lot of effort on both sides to fix. All languages have these issues, but I have put together a preliminary "style and formatting agreement" for my target language, German. The idea behind this agreement is that I will give it to the client for review before the translation is started. Clearly not every client will be capable of making a decision in all these areas, but they can either send it off to the local branch office that can make the appropriate choices or just leave the pre-selected options (which, of course, represent the options that I prefer ;-). Either way there should not be much disagreement after the translation.

And remember, the areas that are covered in such an agreement do not concern errors, but simply stylistic or other preferences. I have uploaded a copy and would be interested in feedback whether something like this is feasible for all kinds of languages.

To be honest, I didn't get much feedback -- so maybe it was a bad idea. But when reader Karin Bauchrowitz reminded me of this last week I could not fail to see some merit in it. If it is relevant, it would be very easy to set up different documents for different languages in Google docs that could be opened up for everyone who is interested in participating.

So let me know. Maybe someone else would like to take on this project. In that case I would be glad to promote it in this newsletter.

. . . who are interested in geeky things like the Okapi Framework takes place next week (Thursday, July 16, 2009 11:00 am PST) and you can register right here.

It truly is not for everyone, but I'm looking forward to it because I have the highest respect for its main developer, Enlaso's Yves Savourel, who will lead this session. This is what the registration materials say about the event and Okapi:

The Okapi Framework, an open-source project, provides a set of cross-platform components and libraries that assist the development of localization and translation-related tools.

The Okapi Framework allows tools developers and localizers to build new localization processes or enhance existing ones to best meet their needs, while preserving compatibility and interoperability. The Framework also provides developers with a way to share -- and re-use -- components across different solutions and promotes existing open standards.

One of the applications built on top of the Framework is Rainbow v6. The application provides an easy way to access many of the functionalities of the Okapi libraries.

In this presentation, Yves Savourel, ENLASO's Solution Architect and active member of the Okapi Framework development team, provides a general overview of the current Framework which includes the system of filters the Okapi Framework provides, as well as the concept of the pipeline and the plug-in steps.

In this session, Savourel illustrates concrete examples on how to use Rainbow and Ratel, two of the tools developed using the Okapi libraries. Some of the tasks demonstrated include:

·         Defining SRX segmentation rules,

·         Preparing documents for translation in different translation environments,

·         Post-processing the translated files,

·         Comparing translations, and more.

I've written about Skype before and how it's become the chosen form of communication in translator communities, albeit primarily as a messaging rather than a telephony tool (I can't even recall when I opened another messaging tool). A new feature that was introduced with the latest version will probably just cement this status. It is now possible to share your desktop with a person you are messaging or talking with.

Well, "sharing" may not be quite the right term since the only thing it does is display your current screen, but this is still very helpful when you have to ask a question about something that does not work quite right or have to show someone how to do it right. At this point it's not possible to have both ends share at the same time or to show something to more than one person at a time; in addition, the refresh rate is just a little slow and the image looks a bit coarse. Though there are lots of other apps that offer this, I STILL like it, if only because it's easier to stay in the program you are in while talking to someone, rather than starting up something different that the other person might have to install first.

If you would like to promote this newsletter by placing a link on your website, I will in turn mention your website in a future edition of the Tool Kit. Just paste the code you find here into the HTML code of your webpage, and the little icon that is displayed on that page with a link to my website will be displayed.

© 2009 International Writers' Group