The Linguist

The Linguist 52,6

The Linguist is a languages magazine for professional linguists, translators, interpreters, language professionals, language teachers, trainers, students and academics with articles on translation, interpreting, business, government, technology

Issue link: https://thelinguist.uberflip.com/i/220935

Contents of this Issue

Navigation

Page 12 of 35

FOCUS: CAT TOOLS highly secretive research and development model, but that the premium version provides the security such companies require. Moses Improving connectivity is a development priority. Coffey said the focus will be on creating a fully interconnected localisation 'ecosystem' that includes project management and accounting systems. Download at globalsight.com TermWiki A terminology management solution that supports collaboration by users anywhere on the web, TermWiki provides access control, user profile management and notifications for changes. It was created more than three years ago by CSOFT, a multinational language service provider headquartered in China. Common Sense Advisory's research has found terminology to be one of the most under-practiced language functions. CSOFT Executive Carl Yao told us that the absence of any major sites devoted to the development, management and sharing of quality terms across languages drove its development. The software is based on MediaWiki – the platform that underpins Wikipedia – and other open-source components. This foundation supports the addition of plug-ins by the developer community. According to CSOFT, the project regularly receives ideas and solutions from its users. TermWiki is a cloud-based service and so doesn't require a download. Users have created more than 5.7 million terms in over 1,700 categories and 102 languages. CSOFT also sells a premium version for businesses that need to customise and integrate its functions with enterprise software. This version competes with commercial products from Acrolinx, Interverbum, SDL and others. TermWiki is best for everyday terminology management. Yao says the public version would not be suitable for companies with a Vol/52 No/6 2013 Until 2006, anyone wanting to use machine translation (MT) had to buy it from a commercial supplier or build it themselves. Then along came the Moses decoder, an open-source statistical MT (SMT) engine birthed by Philipp Koehn and Hieu Hoang, and further enhanced by various universities and institutes. Moses has found a following among corporate users experimenting with MT, language service companies, translation providers building full-service products, and third-party developers creating commercial variants. Some LSPs have built Moses into their workflows; others have incorporated it into web portals that allow anyone to upload TMs and train an MT engine. Given all these configurations, Moses competes with commercial SMT solutions such as Asia Online and SDL BeGlobal. Another open-source MT engine, Apertium, provides a similar training ground for anyone interested in rules-based machine translation. Anyone wanting to experiment with SMT can download Moses directly from Github. However, companies looking for a friendlier Moses application, with basic tools for full production systems, extra-cost hosting and/or consulting, can look to vendors such as Digital Silk Road and Precision Translation Tools. Some offer premium versions as well. Download at www.statmt.org/moses Okapi The Okapi Framework is a Java-based collection of interface specifications, format definitions, components, and applications for localisation projects. Developers can incorporate them in their own projects to create tools for translation and localisation. Two things drove ENLASO to join the project: 1) the commonality of localisation tasks across projects and companies; and 2) a desire to avoid using its competitors' tools. The formal open-source framework enables individual coders to share work, getting the benefit of components created by people who have already conquered the programming demons they face. The Okapi team designed an object model so that the components could communicate with each other, increasing its power and interoperability. Okapi is widely used and its components have been incorporated into in-house tools at Adobe, the Church of Latter Day Saints and VisaTEC. Okapi has also been used by COTS developers, including Asia Online, CrowdIn, MemSource, OmegaT, ONTRAM and Safaba. Downloaders who need premium technical support or customised modules can buy them from ONTRAM and Spartan Consulting. Support is also provided by users and through a developers' mailing list. Anyone interested in developing translation tools or extending the capabilities of other products to deal with localisation issues should find value in Okapi. Of all the solutions mentioned in this article, it has arguably delivered the most value to the widest and most diverse population. Looking ahead, ENLASO's Yves Savourel says performance, scalability, and support for multi-threading and concurrent access will be improved. The team also wants to make Okapi a bit more formal, with contributor license agreements. Download at okapi.opentag.com Given the specific needs of end-users, translation tools are never going to attract the broad developer interest of more mainstream software, such as Firefox, OpenOffice and Red Hat. OSS translation projects will always work with much smaller pools of downloaders and contributors. The problem is gaining enough development momentum outside the small group of original coders or staff donated by a commercial enterprise. That said, these projects are on the same trajectory as innovative products in the commercial sector. The newest revisions have moved to Java, thus allowing for crossplatform use. Web-based TermWiki has bypassed installations altogether. We expect the next generation to rely heavily on the cloud, making the question of which operating system it can run on irrelevant. Finally, we should note that the best OSS projects are characterised by their transparency. Efforts that have a steering committee to guide their development, a strong user community willing to share its views, and a published roadmap, provide a solid foundation for business-critical use. When considering open-source software for long-term or strategic use, look for projects with that clarity of vision. DECEMBER 2013/JANUARY 2014 The Linguist 13

Articles in this issue

Links on this page

Archives of this issue

view archives of The Linguist - The Linguist 52,6