[Isis-users] Inverted file generation for Arabic characters

[De Smet Egbert]

Hello,

with a student from Algeria I did an exercise of using ABCD with Arab records.
It works fine, also indexing and searching, but beware of the following :
- if you are using MARC21 ( you didn't specify which database you use), you have to put the values, e.g. the title, inside a subfield, otherwise it will not be indexed. Now, to put a value inside a subfield, you have to be aware of the fact that Arab goes from right to left, so first you have to put the subfield identifier e.g. ^a, in the Latin script (from left to right), then the value in Arab. I always recommend to use the +-icon in the worksheets to invoke the subfield-editor, avoiding this problem. Maybe that already solves your problem : make sure you put the Arab values into the right subfields in the right way, otherwise they will not appear in the indexes.
- ABCD is not (yet) UNICODE, so you cannot use the UNICODE-codes for Arab characters, we use the HTML-codes instead. In fact the list is given here (I think this is the full 28-characters alphabet) :
ب ت ث ج ح خ د ذ ر ز  ش 
س ص ض ظ ط ع غ  ف ق ك 
ل م ن ه و ى
You enter values just as always after putting the keyboard into Arab mode, but in reality these are the values stored into the record and also indexed.
- doing this previous step will put the indexed values on top of the index as the '&' character preceeds the normal a, b, c, d etc.
So this is not a real solution and we still need UNICODE (a test will be done to find out whether we can do it soon), but maybe for the time being it could serve as a solution.

Egbert de Smet
IOIW / U&S
Universiteit Antwerpen
________________________________________
From: isis-users-bounces at iccisis.org [isis-users-bounces at iccisis.org] on behalf of anwar jabr [anwar_j at yahoo.com]
Sent: Thursday, July 26, 2012 10:39 AM
To: isis-users
Subject: [Isis-users] Inverted file generation for Arabic characters

Dear All,
I have ABCD database with Arabic characters, but when I got to Inverted file generation just the English words are being generated in the index file. Any help?