Myths of Machine Translation

Machine translation has many limitations and I am not a big proponent of complete machine translation but as these systems become more sophisticated, I believe that they do have a place in the translation process.  I read a great article in tcworld on dispelling the myths and bad rap for machine translation. 

Ultimately, most of us have used the free services out there on the net and that is not a true representation of many of the machine translation tools available today.

The free services do not allow users to select a subject field or provide user terminology, they can’t set stylistic preferences or translation parameters other than the specification of desired language pair and the drop in of source text.  Not much to go by.  No wonder there are such problems. Often times what is produced is a word-for-word translation.

The proprietary commercialized machine translation services are usually only utilized in larger organizations.  These systems have cost millions of dollars in research and development. 

Recent facts that dispel some of these myths:

• Machine translation can improve the productivity ands consistency of human translators
• Machine translation enables more translation of materials that would ordinarily not have been translated due to lowered costs and turn around time
• There are many machine translation programs/technologies that are capable of producing translation results in the right environments like rules based machine translation, statistical machine translation and direct machine translation. 
• There are many affordable and user-friendly machine translation packages available for even the smallest of businesses and freelancers.

Machine translation is still a very controversial issue and as a translation agency we will need to utilize and work with this new technology to provide the best value to our customers.  I think there will always be a role for the those with linguistic backgrounds as these systems are meant to be used with human translators.  However, I think in the near future there may be a higher adoption rate of machine translation as more and more technology and sophistication  and solutions are provided to increase efficiency.

The Basics of DITA

DITA – Darwin Information Typing Architecture is the buzz word in technical documentation and translation these days. 

What exactly is DITA?

An information model/content management system that is used for content-rich and multi-channel environments.  According to Wikipedia, DITA is an XML based architecture introduced by IBM in 2001, to automate creation, authoring, producing and delivering technical information. This DTD system divides content into small, self contained topics that can be re-used in different deliverables and reduces information redundancy.

Three Basic Topic Types:

1.  Task – works for procedures and lists of steps to accomplish a task or outcome. 

2.  Concept – contains definitions, rules and guidelines

3.  Reference – describes command syntax, programming instructions, and other reference material.

Features:

• Modular content vs. long book oriented files
• Topics can be reused in various deliverables
• Allows for conditional text, index markers and topic to topic links based on audience, platform, model, and product
• Structure is similar to HTML which can be used directly in DITA Topics
• Topics are easier to find using an extensive metadata
• End to end architecture using specifications of elements, attributes and rules.

DITA can be used in Windows, Linux/UNIX and Mac OS operating systems. 

Output formats:

1. PDF
2. XHTML
3. HTML Help
4. Eclipse help
5. Java
6. Rich Text Format

The DITA Open Toolkit is an active open-source free DTD that has been contributed to from several companies. 

Translation and localization companies like Iterotext Translation Services can use your DTD files and your XML files to provide multi-lingual documentation.

If you are looking for more information please visit the Toolkit User Guide or the DITA Blog.

Translation Software Compatibility

When working with a translation agency or a translator, be sure to ask which versions of your publishing software that they are using and if it is compatible with your versions.

You don’t want to give your projects to a translation company or translator to find the completed translation cannot be read on your computer.  Software updates can be costly, so many free lance translators cannot keep up with the new technology.  Translation agencies and companies typically invest in upgrades to work more seamlessly with their clients. 

It is also important to work on the same platforms of Mac or PC. By having a discussion about software before you hand over your documents, you will ensure what software/version your final project will be in.  A quick tip – this could be easily overcome with a simple PDF, if that is all you need for printing purposes. 

Translation Memory, a useful tool your translators should be using

Today, translation memory (TM) tools are everywhere.  Most translators and translation firms are using TM for contracts, legal documents and technical translation projects. 

25 years ago when the owner of Iterotext designed our own proprietary TM tool, no one else was using this.   We began creating this program due to our client’s to help save time and to increase consistency and ultimately save our clients money.

According wikipedia TM is:

“A translation memory, or TM, is a type of database that is used in software programs designed to aid human translators.  Some software programs that use translation memories are known as translation memory managers (TMM).A translation memory consists of text segments in a source language and their translations into one or more target languages. These segments can be blocks, paragraphs, sentences, or phrases. Individual words are handled by terminology bases and are not within the domain of TM.” 

One of the benefits of TM is that it ensures that the translated documents are consistent, including common definitions, phrasings and terminology.   A translation memory can help accelerating the overall translation process; since translation memories “remember” previously translated material, translators have to translate it only once. Ultimately, a TM Can reducing costs of long-term translation projects; for example the text of manuals, warning messages or series of documents needs to be translated only once and can be used several times.

Typically the segments are broken down into three areas:

New matches – which is a segment that has not been translated before and will be charged at full price.

Fuzzy matches – which is a segment that is typically 80%-99% similar to a segment that was previously translated and is translated at a discount.

100% exact matches – which  is a segment that has been translated before and only needs to be proofread for context.

There are several different programs out there.  Some are free and some you have to pay for.  We use SDLX and Trados which are now owned by the same company.  We use the professional version which has more flexibility and options.  Each translator and agency use the program and service level that is appropriate for their work and budget constraints.  Since Iterotext specializes in the work that requires TM we have spared no expense to ensure that our clients’ projects are timely, consistent and cost effective. 

« Previous Page