Hello everybody, After two night halves of sleeplessness, I am pleased to announce the first solution for an English-Hebrew dictionary for Linux.
What you need: 1. wordtrans sources (from http://wordtrans.sourceforge.net). 2. Babylon's dictionary files, from www.babylon.com. Note that I downloaded mine some time ago, and since then I heard rumors about Babylon changing their shareware policy, so I don't know if it's still freely downloadable from them. Also, as I understand, it's not legal (for me or others) to redistribute them. Comments? Also note, that you download .exe files, but they can be opened with unzip. 3. My patch. It's in here: http://www.cs.tau.ac.il/~didi/wordtrans-heb-diff.gz It's againt 1.0beta2 (from debian sid, few weeks ago), but should apply cleanly to the latest (1.0.2), as the Babylon support files havn't changed for a long time. 4. Whatever that is normally needed to compile wordtrans - g++, several libraries, etc. Look in the site. How to use: untgz, patch, compile. I only tested the "console" version, not the QT/KDE, so you can save a lot of compilation time by doing 'make wordtrans', as that's what I tested. Copy it somewhere, under a different name (I put it as /usr/local/bin/hwordtrans). Make a directory ~/.wordtrans, put configuration files in it. I use debian. It puts some at /etc/wordtrans and when you run it, it insists on copying from /etc to ~ whatever is there, but I think you only need babylon.baby. Find samples in conf/ . Edit babylon.baby to use the hebrew file: Under section '[Diccionarios]', write general = <wherever you have your babylon files>/engtoheb.dic Then do (in an xterm with a hebrew font): hwordtrans <word> | fribidi -charset 8859-8 You can probably also use bidiv of Nadav Har`el. If the QT version works (probably), it might also do bidi. Havn't checked. Some notes: 1. Currently, this patch breaks other languages. I don't know how many of Babylon's languages are supported, probably only Spanish, but it sure breaks some of them. Why? Because the encoding algorithm is completely different (and surprizingly, the Hebrew one is almost the same as the English, and different from the 'target' (probably only Spanish currently)), and because I havn't yet thought of how to decide which one to use. So it only uses hebrew (with the patch applied). I havn't yet contacted the developers of wordtrans (or babytrans, on which wordtrans' Babylon support is based), so I don't know much currently. 2. Theoretically, you might get a few 'X' chars in the output. I havn't, but it can happen. If you do, check to see what the real Babylon tells you. It should be some special character. If it is, tough. I don't think it would be easy to make a font that will give the exact same results. If it isn't, it's a bug. Please let me know (email the source word). 3. If interested, look at the source. It's very similar to the English encoding. Of course, I didn't know that, so I spent many hours with a binary editor and babylon, to see how it reacts. 4. Please try it and report success or problems. 5. If anyone has an opinion/idea about how to merge it with upstream, without breaking Spanish, please share it. I am sorry for the crosspost and the long email, but you can imagine I was very excited when it worked (an hour ago). Also note, that I CC Ricardo Villalba (who wrote wordtrans) and Frederic Jolliton (who wrote babytrans), whom I of course thank a lot, BTW, Thanks!, so think before you do a 'g'roup reply. Didi ================================================================= To unsubscribe, send mail to [EMAIL PROTECTED] with the word "unsubscribe" in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
