Re: Clang/LLVM migration roadmap

[Riccardo Mottola]

Hi Daniel,

I think unvoluntary you made a very important point, I don't know by being 
aware of it or not.


> If you’re looking to recruit more GNUStep users, I just don’t think you 
> can do that without ARC, @[], @{}, etc. There are plenty of ObjC developers 
> out there right now who are potential future GNUSteppers. And as you recruit 
> people, I think “you have to use ObjC instead of Swift” is a much easier 
> pitch than “you have to do manual memory management and type out 
> objectAtIndex: every time you want to get an object from an array”

You used the term "user" and that's wonderful because it made me think about 
it. I agree and always thought we need more users.

But what is a user? I can think at least of two types.

1) User of GNUstep libraries because user of GNUstep applications. This kind of 
person doesn't care about runtimes and memory management, he just installs 
applications by some means and uses them. He cares about stability, speed, easy 
of installation. That's the most common user, he consumes GNUstep... like a 
user of GIMP not knowing it is based on GTK. Or the Mac being based on Cocoa, 
etc.
2) User of GNUstep because he writes GNUstep applications, for work, pleasure; 
in a mix of existing apps or not. That person may care what kind of language 
and runtime he needs, but still he doesn't care what GNUstep core (and other 
applications he perhaps uses, like GWorkspace, ProjectCenter or GORM) is 
written in. He only needs to be able to use ARC, properties (or not, if he 
chooses not to). So he needs Clang and libobjc2 as for now if he needs ARC. The 
only thing he needs to worry is that the core libraries need to be compiled 
with ARC support and that. ARC can be turned on for module and not another, 
David teaches us

Only a contributor to GNUstep itself needs to worry about what language it is 
written and what code rules it has.

Of course, every user kind of 1) and 2) may become a contributor; but I deem it 
prefectly acceptable to give him freedom to code whatever he wants, but for 
GNUstep "tiself" additional restrictions can be applied.

I always did know these disctinctions, but it makes me see this discussion with 
a new mindset -à and take the hype and urgency out of it.

Still, libobjc2 can benefit a better support, integration, ease of install, 
platform support. But the urge of migration gets relative.

Riccardo

-- 
Proudly sent with GNUMail on GNUstep on FreeBSD/x86-64