11–20 of 141 posts
Meanwhile, in the free software world... https://docs.google.com/presentation/d/1ySTZaXP5XKFg0OpmHZM0... https://github.com/behdad/harfbuzz/blob/master/test/shaping/... (Harfbuzz is the default text layout engine for Linux GUI frameworks: Qt, GTK, etc)
> Utility had defeated tradition. Is that supposed to be self-evidently bad?
That was an amazing article, and a great insight into middle eastern scripts in the digital world. It doesn't look like it'll get out of the new queue, but I wanted to thank OP for sharing it anyway.
There are arabic derived scripts at least as far east as China and the Philippines (jawi), but describing them as Middle Eastern is plain wrong - just as you would not describe Balinese or Cambodian as Indus Valley.
For web development, perhaps you could create a simple system that detects if the user's device supports nastaliq. If not convert, it to naskh or Romanized Urdu (not sure which is preferred). As for dealing with all the data already written in naskh and Romanized Urdu, this is a bit more complicated because I am guessing there is not a one to one match between naskh and nastaliq (naskh has less letters), and Romanize…
Data is stored as Unicode (except some really old desktop app formats that were written before Unicode). Naskh and Nastaliq are font styles, used when the system displays the data. The solution is simply that good Nastaliq fonts be created and then used on websites where Urdu is being used. Since Urdu fonts can be much larger in size than Roman fonts (the fonts have to specify joining behavior between all possible co…
My assumption is that not all OSes will adopt Nastaliq simultaneously. So assume that the website is storing data in complete coding for Nastaliq and they want to send it to a device that cannot render Nastaliq so the Unicode should be converted so it is properly renderable in Naskh.
The other half is just to make one's Urdu language web experience completely in Nastaliq. Even unimportant stuff written in Roman Urdu and then converted to Nastaliq might help promote the use of Nastaliq.
Earlier quoted context omitted.
Data is stored as Unicode (except some really old desktop app formats that were written before Unicode). Naskh and Nastaliq are font styles, used when the system displays the data. The solution is simply that good Nastaliq fonts be created and then used on websites where Urdu is being used. Since Urdu fonts can be much larger in size than Roman fonts (the fonts have to specify joining behavior between all possible co…
The author of the article mentions that Naskh does not contain all the letters that Nastaliq has, so if someone types an Urdu text in Naskh it would not have the same underlying Unicode if they had typed it in Nastaliq. Correct me if this is wrong. My assumption is that not all OSes will adopt Nastaliq simultaneously. So assume that the website is storing data in complete coding for Nastaliq and they want to send it…
Both Naskh and Nastaliq can be encoded using the Arabic block in Unicode, it's just that Nastaliq's vertically stacking nature makes it difficult to deal with in computer systems that expect clean rows of text.
Such fonts have already been designed. The issue now is to find Android and iOS developers who can make them default in the operating systems.
When I talked to Microsoft they said that on the smaller devices every last bit of memory was precious so they didn't want to stick extra megabytes or however big a font is. I think that Apple's cheapening out is worse than Android's, because at least Android offers the entire Urdu alphabet (though in naskh). Apple doesn't even offer the Urdu alphabet, requiring Urdu users to have to make do with 12 less letters.
I have a few thoughts on this though. One would be the legibility. Of course I cannot judge legibility of writing systems I do not like, but it seems that nastaliq would be hardly readable on a lot of mobile devices and I wonder how difficult learning the ornate script is. I am talking about alphabetization here.
Next thing is: I am learning turkish and turkish is written in the roman alphabet. As far as I know, it was written in a arabic/persian script before which was then reformed to use the latin alphabet. As far as I can tell this is today really uncontroversial and using the latin alphabet is actually the more suitable alphabet for turkish and its rich vowel system that is really important for grammar and meaning. Again I cannot really say anything for Urdu, but knowing it is not arabic but afaik a language of the indo-european family I wonder if there are more reasons to use the lating alphabet than just availability of nataliq fonts and rendering engines.
As a side note I would add a few observations relating to the cultural/heritage aspects. In Germany, the "Fraktur" was used widely even at the beginning of the 20th century (see http://en.wikipedia.org/wiki/Fraktur ). Some authors, like Hermann Hesse refused that Antiqua fonts would be used for their writings until publishers convinced them that their works could just not be read by young folks. In a way, a lot of people argued against using non-gothic fonts, but in the end antiqua became quite standard. Nowadays we use the lating alphabet (and most people are not concious about that there ever was a switch).
> Utility had defeated tradition. Is that supposed to be self-evidently bad?
Indeed. I feel like languages/scripts dying out has some advantages too.