Hi Richard,
there are two different issues, on of them is the implementation of Arabic shaping algorithm and the other one is the implementation of bidirectional algorithm based on the W3C recommendations for XSL. http://www.w3.org/TR/xsl/#writing-mode http://www.w3.org/TR/xsl/#d0e4879 [EMAIL PROTECTED] wrote: > Hi, > > At some point in the past you did some work on getting Apache FOP > to work in Arabic: http://www.anneundsebp.de/fop/fop.html > yes, Sebastian has also word on it. > I'd like to see fully working Arabic (and Farsi) support for FOP > working and am willing to devote some time to getting it done but > don't want to duplicate anyone else's work if I don't have to. > > With this in mind I have a few questions which I'd be very grateful > if you could answer: > 1. What is the state of this work? > I got some ideas from other projects and could manged it to get the correct Arabic shaping. The Arabic shaping algorithm is implanted now. The bidi implementation is more complex and need more time to be implemented based on the W3C recommendation. The most important part is to implement the writing-mod property which can get the following values: / /lr-tb | rl-tb | tb-rl | tb-lr | bt-lr | bt-rl | lr-bt | rl-bt | lr-alternating-rl-bt | lr-alternating-rl-tb | lr-inverting-rl-bt | lr-inverting-rl-tb | tb-lr-in-lr-pairs | lr | rl | tb | inherit the implementation for the rt-tb and lt-tb is already done about 60%, but the implementation needs refactoring and should be checked for the conformations with line breaking algorithm. It works fine with first line but not with a complete paragraph. I don't know why, should just read more about Area trees in FOP. The direction property could be implemented parallel to writing mode. The property inheritance should be also implemented. I think the last part is to implement the unicode-bidi, which is not very discommoding when it is not implemented. > 2. Are you continuing to develop it? > yes, but I don't have much time for them. I just work on weekends. > 3. Against which version was the current fork made? > the last trunk version > 4. Why doesn't the link to the binaries work? > the binary file was very buggy, it could not connect the Arabic characters correctly and could not handle the Arabic-indic numbers. This link is from sebastian weber, and he decided to remove the binary files. > 5. Are you intending to submit a patch against the current FOP > trunk? > If you wish, I can send you this patch via email. Don't wanted to apply an unfinished implementation which would be rejected. The other issue was that I need an Unicode TTF-font with Apache license which implements all of the Arabic characters and could be used for test cases. Fonts with GPL license could not be accepted. > 6. If not, could this code be used as a basis for any future > update? > yes, I try to finish at least the write-mod so that it could be used in other projects. I need to use it with DocBook, it could be enough for me when it works correctly with rl-tb and inheritance. > Regards, > > Richard > >