Meedan

The Meedan Blog Archive

How to use MeedanMemory

So, now that the MeedanMemory became available, how can you make best use of it?

First, lets clear some basics. MeedanMemory is a translation memory that is focused on Arabic and English. This memory contains text fragments (sentences) in aligned pairs of Arabic and English. These sentence pairs, also called translation memory segments, can be used in different ways. We will focus here on using them for translation purposes.

Translators can use these segments, previously translated by other translators, during their work to find similar or approximate matches, and to help define the translation context of certain sentences. So for example, if the source text contains the sentence “The door is red”, the translation memory will be queried using fuzzy matching for exact or similar translations. If the translation memory contains a similar translation for the sentence “The door is green”, this sentence can be presented to the translator, who can then modify the sentence, or better understand the translation context. This is particularly helpful in cases where sentences are frequently repeated, such as software localization and translation of instruction manuals.

Translation memories are usually stored in an industry standard file format called Translation Memory Exchange (TMX). Using this standard facilitates the integration of translation memories across different software applications and platforms. MeedanMemory is packaged as a standard-compliant TMX, and can be downloaded from the following location: http://github.com/anastaw/Meedan-Memory.

So, to start using MeedanMemory, you need first to download the TMX file. Then you will need a software application that understand this format and is capable of performing the required search and analysis functions. These applications are usually called translation memory managers, or computer assisted translation (CAT) software. You can find a list of possible options here. There are several open source and proprietary alternatives. OmegaT is a widely used and actively developed possible option.

After downloading and installing OmegaT, you will need to configure it to use MeedanMemory. From the “Project” menu, select “New”. The Create New Project dialog will appear.

In the “Translation Memory Directory” section, select the location that contains the downloaded MeedanMemory file. Click on “OK” and you should be able to start using MeedanMemory right away. Import the source file that needs to be translated by choosing “Project -> Import Source Files”. OmegaT will analyze the file during the import, break it down into text segments, and display the similar translation matches for each segment.

We hope that this brief introduction can get you started in using MeedanMemory. You will certainly need to learn more about the software application you will decide to use. The application’s website and user manual are excellent places to start. We are sure that once you get used to the features and controls of these applications, you will quickly appreciate the benefits an open access translation memory, such as MeedanMemory, can provide.

Comments on this post

2009-11-12 04:41:44 -0800
hmm. i did all this and it all seemed to load ok. but there are two issues: 1) it seems to take a long time to load the memory tmx file - so much so that omegat seems to freeze at times and 2) then i'm not getting any suggestions in the memory windows. i'm testing this by using the wikipedia english article on hamas: http://en.wikipedia.org/wiki/Hamas which is pretty long so i'm surprised the large memory doesn't offer any suggestions. i've tried skipping backwards and forwards through the sections but nothing happens. any advice?
2009-11-12 04:46:32 -0800
Thanks for feedback! We'll try to get a full response to this shortly, though the team is travelling today and tomorrow.
2009-12-22 08:06:12 -0800
Hello every one , and thanx for the nice work .. i work with SMT and i wonder if Any ideas how can use Madeen with Moses ? or how can we extract TMX files into a two raw text files (English , Arabic )
2009-12-24 15:22:19 -0800
Hi Mossab, It is certainly possible to extract the content from the TMX file into a text file. You can try to use Olifan: http://okapi.sourceforge.net/Release/Olifant/Help/ which will enable you to export the TMX file content into a text file. I hope that helps
2009-12-25 10:14:06 -0800
Thanx alot brother .. you saved me from a daunting Perl programming task i was going to do to fullfill this task .. the software works just great ..
2009-12-30 08:17:27 -0800
Hello Again every one , i am trying to add Dictionary for my Statistical Machine Translator , and i am searching for English Arabic dictionaries in a basic raw .txt or .xls format .. any one have any idea where to find such a file ?? Again thanx for ur support Mossab
2009-12-30 16:17:27 -0800
Hello Again every one , i am trying to add Dictionary for my Statistical Machine Translator , and i am searching for English Arabic dictionaries in a basic raw .txt or .xls format .. any one have any idea where to find such a file ?? Again thanx for ur support Mossab
2010-06-26 11:30:06 -0700
would u type in English if u want us to Answer you :)