My biggest concern with trying to do a book right now is that the language hasn’t hit 1.0 yet, and the docs for core parts of the language are…spartan. It’s a deliberate decision (and one I agree with) to hold off on too much polish until the language has firmed up more, but without the official docs being in that place it seems to me that any books published will either harm end users (because hey, the book becomes obsolete) or the language (because they suddenly feel pressure to hang onto whatever legacy stuff the book taught to new developers).
That’s the same recommendation I give in the blog post: avoid having a publisher tempt you into writing a complete Zig manual.
That said there are plenty of people who struggle with systems programming concepts and up until now all book recommendations I’ve seen were too tied to C-isms and x86_64-isms, which is not good since Zig can be a way of learning lower-level programming concepts without having to face the full blunt of C cruft (it certainly was for me).
In the book that I’m working on I’m not going to teach people Zig past the basics, but I will instead focus on things like stack vs heap, what’s a syscall, etc.
I think there is space for plenty of books of this kind that focus on computer architecture basics but without assuming that the way C / x86_64 does things is a universal truth.
or the language (because they suddenly feel pressure to hang onto whatever legacy stuff the book taught to new developers)
That’s not a thing in our case, regardless of how many books might get published :^)
I agree. Working with Zig this week, I already encountered two changes that are in the nightly version that differ from 0.10.1: @memset being different and enumerating being possible for iterating. The language is not stable enough for a printed book right now.
This was my thought as well … and in general I am typically hesitant about hard-copy books for a language because by the time they go through the editing & publishing process, a lot of things can fall out of date.
Off-topic: something like the rust book which is available online and constantly updated would be wonderful. Even better if it’s open to contribution for corrections and additions.
My biggest concern with trying to do a book right now is that the language hasn’t hit 1.0 yet, and the docs for core parts of the language are…spartan. It’s a deliberate decision (and one I agree with) to hold off on too much polish until the language has firmed up more, but without the official docs being in that place it seems to me that any books published will either harm end users (because hey, the book becomes obsolete) or the language (because they suddenly feel pressure to hang onto whatever legacy stuff the book taught to new developers).
That’s the same recommendation I give in the blog post: avoid having a publisher tempt you into writing a complete Zig manual.
That said there are plenty of people who struggle with systems programming concepts and up until now all book recommendations I’ve seen were too tied to C-isms and x86_64-isms, which is not good since Zig can be a way of learning lower-level programming concepts without having to face the full blunt of C cruft (it certainly was for me).
In the book that I’m working on I’m not going to teach people Zig past the basics, but I will instead focus on things like stack vs heap, what’s a syscall, etc.
I think there is space for plenty of books of this kind that focus on computer architecture basics but without assuming that the way C / x86_64 does things is a universal truth.
That’s not a thing in our case, regardless of how many books might get published :^)
I agree. Working with Zig this week, I already encountered two changes that are in the nightly version that differ from 0.10.1: @memset being different and enumerating being possible for iterating. The language is not stable enough for a printed book right now.
This was my thought as well … and in general I am typically hesitant about hard-copy books for a language because by the time they go through the editing & publishing process, a lot of things can fall out of date.
Off-topic: something like the rust book which is available online and constantly updated would be wonderful. Even better if it’s open to contribution for corrections and additions.