Thanks for positive comments from Khaled and Afief, and even fixes -- that's fast turnaround! I would like to clarify my previous post and maybe suggest one or two things.
-- One would > hope that the work with the ligature tables could be re-used by the > individual font designer, but there are traps. For instance, one font with > very long fathas has them running into the next letter. > I'm not sure what you are talking about here, as Arabeyes fonts in general has very few ligatures, may be you mean substitution tables for initial/medial/final forms? Answer: Bad wording on my part. The big thing with the August release, as I read it, was taking into account what comes after a letter with diacritics. I assumed this has been done in the same way in Uniscribe/Unifont as with ligatures, e.g. lam + aleef. In other words, a (growing) table of most combinations of letter + diacritic(s) + following letter. Let's call them bigram (with diacritics) tables. >* fatha over shaddah are too far apart -- Take this to means that the shadda is much too far from the letter itself. -- I do not understand what you mean with "fatha shadda fatha *followed by* tha fatha" but I asume it's simply a case of too many diacritics running into one another, this is hard to fix as it needs lots of checking with individual letters... Answer: I meant like in this word, the placement of the shadda causes it to almost run into the three dots of the thaa. (Maybe not in the font this word is rendered, but several others.) وَحَدَّثَنِي The simplest fix which probably would look much nicer is to lower the shadda and possibly move it a little to the right, if this does not contradict Arabic typography. It may not solve the problem of shadda with kasra below, since obviously the shadda has to be higher then. Move it to the right. A lot of work has been going into the fonts recently, but unfortunately the people who work on the fonts are not necessarily the people who use the fonts on a daily basis -- You point to a solution: If one prepared a reasonably big sample of varied text, it could be rendered in the font to be checked. Any experienced Arabic reader would incredibly quickly spot placement errors. This would be a Monte Carlo method, as there would be no guarantee of discovery of all "bugs". A more boring method would be to have a sample with all bigrams with diacritics. One would assume that there is feedback to the original font designer he/she ought to have a big interest in checking these improvements. Surely we all can come up with more improvements. Best for now, Goran _______________________________________________ Developer mailing list [email protected] http://lists.arabeyes.org/mailman/listinfo/developer
