07 Nov 2009
Today Meedan is releasing the first publicly available open licensed Arabic/English Translation Memory (TM). This release comes as a part of Meedan’s ongoing commitment to improving the state of Arabic language machine processing research and supporting translators and machine translation researchers.
The new translation memory, called MeedanMemory, contains around one million words in aligned Arabic and English sentence pairs. It constitutes an invaluable resource for translation professionals, Arabic linguistic researchers, academics and volunteer communities, and is the first translation memory of such size and sophistication to be released under an open access license.
MeedanMemory is packaged as a standards-compliant Translation Memory Exchange (TMX) file, which facilitates its integration with most computer assisted translation tools and translation memory managers.
Professional translators can utilize MeedanMemory to support their translation work and improving its efficiency, and to increase consistency and quality in the translation outcome. Researchers can use this resource in designing and implementing their research activities into Arabic computer linguistics and machine language processing, which will lead to the development of better text search engines and document analysis tools and applications. And community organizations, non-profits and volunteer communities can leverage MeedanMemory to facilitate translation between Arabic and English of the content they produce for lower costs, lower turnaround times and higher quality.
The MeedanMemory is released under the Open Database License, which permits the use, copy and distribution of the translation memory, produce work using it, and to modify, transform and build upon the translation memory.
Meedan is pleased to announce this first release of MeedanMemory, and will continue to release data generated from the Meedan translation community into this open repository. We also hope that other organizations will contribute to the growth of this unique and valuable resource for the Arabic language and Arabic digital content that holds great potential for supporting Arabic content projects and initiatives.
To download your copy of MeedanMemory, please visit: http://github.com/anastaw/Meedan-Memory
UPDATE (2013): After a user reported the above link broken, please use the following URL to access MeedanMemory: https://github.com/meedan/Meedan-Memory
For more information about MeedanMemory, please contact: firstname.lastname@example.org Meedan Cross-language Arabic-English Social Networking: beta.meedan.net