On Thu, 2017-06-29 at 07:09 -0700, Kerim Aydin wrote: > But hey - this Arabic stuff?? Well, it's not some important language, > like say Japanese. Let's just translate it to bytes and ignore the > meaning, eh? Completely re-arrange the word order like no native > speaker, and not even a translation machine, would do, eh? I guess > that's fine. Basic principles of reading with good faith don't apply > to a language like *that*. Let's talk about byte order, instead.
Sorry for harping on about this, I'm just really annoyed myself at what you've written. It feels like all the effort I've gone to to do things like understand and work with languages like Arabic have been wasted. A lot of people have put in a lot of effort in order to create communication standards that allow text in all languages, not just English, to be understood unambiguously when communicated from one person to another. And now I'm finding out that that all that work is irrelevant, because when people actually write in Arabic, I'm expected to ignore what what they say actually means, and assume that I should take the primitive understanding that it's all just left-to-right, left-margin-justified text? Arabic has its own rules for writing it, whether on paper or on computer. (The very simplest is that, whether on paper or computer, you start at the right hand side of the page.) If you don't follow those rules, it shouldn't be surprising that the meaning that people ascribe to the message you send isn't the same as the one you intended. In particular, following the same rules as for English is going to produce a result that's meaningless in Arabic; text's going to wrap in the wrong places, embedded quotations in left-to-right languages will be in the wrong places, and so on. It can, however, sometimes produce a result that's meaningful in English, especially when there's English text in the same sentence. (To be honest, I was expecting that you'd follow up your CFJ by submitting the same thing as an image, which can't be reflowed or parsed and which is therefore missing the context you'd need to be able to unambiguously determine the direction it was written in. I was surprised by the apparent lack of understanding of encoding standards for writing various different languages. Perhaps this is the fault of computer software generally still being rather English-centric, and making entering text in other languages more error-prone than it should be; I know I've seen my email client produce incorrect or suboptimal results both with your Arabic, and with 天火狐's Japanese. This is something I'm working on at the moment – I'm trying to write a rendering library which handles all these languages correctly.) -- ais523