Advanced xml techniques
Extensive use is made of advanced xml techniques to overlay a linguistic interpretation onto your text data. This overlay is not obtrusive.
Author memory - control over the source language material
We concentrate our technology on the source language text and are able to monitor changes at the sentence level between revisions of a document. Control over the source language is an important cornerstone to reducing translation costs.
DOM Differencing
Using author memory we are able to accurately work out what has changed at the sentence level within a document following a revision. DOM stands for Document Object Model. We use the hierarchical DOM structure to work out what changes in the text occurred between revisions of a document. DOM differencing is highly beneficial to reducing translation costs.
Perfect Matching® - reducing translation costs
Using the previously described techniques we can guarantee that, for text that has not changed between updates, exactly the same target language translation may be used. This means that the translation does not need to be reviewed by a translator. This is an important distinction compared with the type of leveraged translation memory process offered by translation agencies. With leveraged matching a translator still has to review the matches to check for correct grammatical context. This adds to the costs.
Advanced Linguistic Matching and Web Search Technology
We use advanced text searching techniques based on Bayesian belief networks and encompassing linguistic techniques such as morphological reduction. This is the same technology that is used by advanced web search engines. Most of the existing translation memory systems use very primitive technology in comparison, and produce poorer results.
Advanced Linguistic Terminology Management
XTM Terminology is fully integrated with the system. When preparing new text for translation a linguistic analysis of the source material is made. Noun phrases are analyzed and compared against the terms in the database. Any matches are highlighted in XTM Editor and the approved target language terms offered to translators.
Linguistic analysis of text units to help reduce word counts
Where text segments contain non-translatable data such as numeric, alpha-numeric, part number, punctuation or measurement data, then the text from these elements will not be included in the word count. Many CAT Tools will include such data in word counts increasing the costs to the customer. Such elements can be protected from translation within XTM.
All the necessary data is prepared for the translators
All the necessary data including leveraged, fuzzy matched and terminology data is extracted and put into a special 'package' ready for translation. We use the OpenTag format as the basis for text extraction. This package can be used to provide on-line web access to the translator directly from your web site. Translation, review and publishing can all be undertaken from this interface, substantially reducing turn-around times. There are no middlemen to slow down the process and add to costs. All word counts are automatically produced as part of the process.
Provide translators with an on-line web browser interface directly from your web site
XTM Editor presents the translator with an on-line translation interface via the internet. The translator can key in the translations directly, with full access to terminology and leveraged and fuzzy matched data. The same interface can be used by the translator to have a formatted view of the text as well as for review and QA purposes. The data can also be merged automatically using the same interface to create the new target language version of the data. This can significantly reduce turn-around times. The translated data can also be reviewed and QA'ed on-line as well. We also provide translators with a formatted preview of the material when required.
Leveraging existing standards
We utilize existing standards whenever possible. Thus xml:tm uses the LISA OSCAR SRX standard for segmentation and OASIS XLIFF 1.0
Supported Platforms:
We have designed our software to run on Solaris, HPUX, Windows NT/2000/XP/2003 and Linux. XTM is written entirely in Java and will run on any platform that has is supported by Java. XTM requires a relational database to be installed such as Oracle, MS SQL Server, PostgreSQL or MySQL.
