|
Home / Overview / New Rules for Translation Requirements
Very few
organizations are staffed with the specialized-software engineers,
linguists and translators needed to conduct a successful multilingual
project.
Costs aside,
for information-based businesses, human translation is simply too slow
and too costly to keep up with ever changing data. Time-sensitive
applications such as news monitoring or stock trading can become
obsolete in the hours, or even days that are needed for human
translation. Multilingual customer support and communication also
require on-demand translation, and often involve large volumes of text.
Some examples are Internet self-help applications such as self-service
technical support, email support and click-to-chat support. Users of
these applications expect immediate results, a standard that would be
impossible to meet with human translation. Even when translations are
not needed immediately, it is prohibitively expensive to translate
high-volume customer communications. Likewise, within organizations,
email and instant messages require rapid turnaround at high volumes.
These cannot be addressed with human translation solutions. Human
translation and localization services are not feasible for
highly-dynamic content. Many software companies maintain large
knowledge bases of support materials. This content changes
substantially as new bug reports, fixes and
upgrade information are added. The cost and delays of retranslating a large, dynamic content source is
unsupportable for any business. Some companies have attempted to use
Translation Memory systems in conjunction with
content management software to manage translation of their content.
This solution is not adequate for dynamic content. TM
systems rely on an existing resource of bilingual translation data. For
example, a Translation Memory system can store a set of recurring
sentences or phrases, and the translations that were made. Later, these
translations can be simply “plugged in”, speeding the
translation process.
For dynamic
content, however, the percentage of reused content is low, making Translation Memory systems of limited value.
Further, a Translation Memory system is not useful until it contains a
large database of parallel source and target text. Producing the source
and target text requires substantial effort by human translators. The
cost and delay of human translation is unsupportable for any high
volume or dynamic content. Automation of the translation process with
Automated Real-Time Translation Software is the only viable solution
for the majority of these applications. To date, companies have
implemented this technology because of the compelling return on
investment of Automated Real-Time Translation. In some cases also,
organized opposition by translators has been a factor in the decision
not to use Automated Real-Time Translation. The quality issue is
central to the debate, and is the most frequently cited reason for not
automating translation functions. This contention is addressed on
multiple fronts, through integration support, custom lexical
development, and quality review tools and metrics. The Mesa Group works
with customers to ensure they understand the Automated Real-Time
Translation workflow. This is critical to ensuring a successful
implementation of Automated Real-Time Translation.
|
