Klaus Lagally wrote: > Dear all, > > I have been playing around with the sources of the dictionary (thank > you, Djihed Afifi, for granting me access), and I also developed some > software to maintain and reformat it. I have no write access to the > Arabeyes servers (and do not want it), so some member of the project > will have to collect the material if interested. I put the material on > my own server for free downloading, and I waive all owner's rights on it. > > At the URL <129.69.218.213/ifi/bs/research/arabdict/> you will find: > > - a file 'techdict.pdf' containing a new, two-column version of the > dictionary, based on the data of April 29, 2007. > > - a file 'arabdict.pdf' based on the same data, but formatted right- > to-left and sorted according to the Arabic collating sequence, for > Arabic/English lookup. Multiple Arabic translations for the same > English term produce multiple entries; terms without any translation > are omitted (what could we do with them otherwise?). > > - a file 'makedict.tgz' containing the software to recreate the above > files whenever new input data become available. There is a Perl script > to perform and control the processing task, plus a rather large number > of utility files. The presence of 'pdflatex' and of a reasonably current > version of ArabTeX are assumed. For more info see 'readme', 'manifest', > and the source files themselves; there are comments to explain what is > expected to happen, as the code is probably unreadable except for some > TeX experts :-) > > Note: The input data are expected to be in UTF/8 encoding in the format > English text====Arabic text > The English text is supposed to contain only ASCII characters from the > range U+0000 - U+007F, whereas the Arabic text may only contain characters > from the Arabic block U+0600 - U+06FF, the Lam-Alif ligature glyphs from > U+FEF5 - U+FEFC, and the punctuation marks ONLY from the ASCII range > U+0000 - U+007F. Any other ASCII text and any other symbols outside the > indicated ranges may produce garbage or might even blow up the processing. > These restrictions come from the fact that the Arabic field is parsed in > order to compute a sorting key, and the parser expects Arabic only. > > There are, however, two exceptions: > > - In the Arabic field, references to other English words are possible > in the format 'see ASCII text' (without the quotes!). > > - Within the Arabic field, English words may appear if and only if they > are included in <angle brackets>. > > (This info ought to reside in the 'readme' file, and will migrate there; > I leave it here to give the readers a hint on what is happening below > the surface). > > In case of problems, bugs, or just strange effects please complain by > private E-Mail. > > Peace > > Klaus > Good Work! Thanks for your efforts!
Mohamed Magdy _______________________________________________ General mailing list [email protected] http://lists.arabeyes.org/mailman/listinfo/general
