Welocalize’s translation management system (TMS) GlobalSight has been around for more than 15 years. Prior to 2005, it was an off-the-shelf commercial offering. In 2008, Welocalize, a global language services provider, acquired Transware which owned GlobalSight software as an in-house tool.
Making GlobalSight open source in 2009 was a business decision by Welocalize, as it allowed users and clients the most options to support and create solutions that work best for them. As it turned out, clients liked the decision and Welocalize embraced the open source model as a business strategy. The GlobalSight community has been active since then and is a vibrant, active group of users, developers, and translation professionals. Users like GlobalSight because it is a fully featured TMS system, which is core to supporting localization and translation programs in large enterprises.
TMS implementations manage the many tasks of human language translation. They enable the translation and localization workflow to be more productive, getting multilingual content to market faster. An open source TMS cuts out all the human touch-points because everything is centralized and files flow automatically, which significantly increases efficiency and saves valuable project manager time.
As an enterprise-level TMS, GlobalSight encompasses workflows, translation, review, translation memory, file management, and reporting. GlobalSight is currently implemented on RedHat JBoss community version 7.2 (EAP 6) with MySQL version 5.5 as the backend database. Under the patronage of Welocalize, the GlobalSight system has undergone a steady process of upgrades under the hood, focusing on performance and stability. The API set has been extended and modified, with focus on connection to content management systems (CMS), machine translation (MT) systems, as well as interconnectivity to Welocalize internal business systems.
TMS’s such as GlobalSight manage the basic task of extracting content for translation into sentences, leveraging those sentences against the translation memory (TM). It moves groups of sentences around from language services provider to provider and delivering the translated content back to the client. The basic problem to solve hasn’t changed that much over the years—each solution provider has a feature set to manage the basic task. The practical problems of different file types, extraction techniques, and exactly what a sentence should consist of is still the same problem faced, no matter what TMS is being used.
While being open source, GlobalSight isn’t free software. Deploying an enterprise-level software platform requires planning, testing, expertise, Internet bandwidth, and hardware to run on. Like any enterprise level software, other costs still need to be counted, even though the licensing costs are zero. We recognize that adhering to international standards, such as XLIFF 1.2 and TMX 1.4b, allows greater interoperability with other tools whom also adhere to the same standards. This helps to avoid proprietary tool lock-in.
As GlobalSight has a large code base, close to 1.5 million lines of code, we don’t get many folks contributing to the code on a day-to-day basis today. What usually happens is that an organization looking to enhance GlobalSight will collaborate with our own devleopment team to ensure we minimise any code forking. Other times, Welocalize will get asked to implement a particular feature and we cost the development time back to the requester. Having the code and expertise at our fingertips allows wider innovation when finding solutions for Welocalize clients. We do sell support contracts. Where GlobalSight fits into the Welocalize business strategy is to add value, technical innovation and leverage to client’s total translation programs to provide greater efficiency and productivity.
The GlobalSight development team at Welocalize is driven to look for ways to improving the translator and reviewer experience. The review stage in the translation process has its challenges as reviewers tend not to want tools that are unfamiliar or take time to learn how to use, yet they want a preview of the final document and be able to comment on it.
For example, while GlobalSight can provide an Excel document with the sentences to comment upon, preview and comments still don’t mix well. Translators want a streamlined user expense. Welocalize recently collaborated with localization industry experts to create the open source translation tool that works with GlobalSight called OmegaT. OmegaT is a free, open source, Java-based CAT application for professional translators. Working as part of this project helps us to further understand translation behaviour, speed and quality. It is a commitment to continuous improvements.
Machine translation (MT) and language automation play a significant role in localization and translation in the commercial world. GlobalSight has integrated various MT engines into the TMS to leverage, translation and reporting processes.
We are also building in new industry standards and developing connectors to other content management systems. GlobalSight has multiple connectors to allow for the integration of content via multiple technology sources. Welocalize recently developed a new COTI connector for GlobalSight. COTI stands for Common Translation Interface and was established by DERCOM, the Association of German Manufacturers of Authoring and Content Management Systems. GlobalSight connects with most large-scale content management systems today.
You can find the source code for GlobalSight on Sourceforge.
Welocalize maintain a global team of support specialists for GlobalSight. The development and QA teams are based at Welocalize offices in China—some developers have been working on the code for many years. Welocalize makes regular public releases of the latest version of GlobalSight, often with new features and functionality, therefore Sourceforge is updated several times a year, often with new features and functionality. The latest release was in October 2014.
Comments are closed.