First of all, thanks for responding. I understand the hesitation of language maintainers with respect to adding complex modules like JSON and YAML, but I do think that there are some major benefits to doing so:
* Users don't have to worry about inconsistent distributions. Following a test suite is one thing, but test suites are (almost) never comprehensive and will not catch programmer errors with respect to particular systems (e.g, LP64 and LLP64). Everything will work until it doesn't, and programmers will have to provide edge-case workarounds until individual distributions fix their implementations.
The same goes for module managers. It's good to provide an open spec and encourage multiple frontends, but it's also good to provide a uniform interface for system administrators who just want to automate module installation for an application or framework. Python's easy_install wasn't ideal, but it provided exactly that sort of common denominator.
"Batteries included" makes it sound like I want my language to come with a playground, which I think is an unfair way to characterize the expectation that Perl 6 provide feature parity with the languages it competes with (i.e., Ruby 2 and Python 3).
* Users don't have to worry about inconsistent distributions. Following a test suite is one thing, but test suites are (almost) never comprehensive and will not catch programmer errors with respect to particular systems (e.g, LP64 and LLP64). Everything will work until it doesn't, and programmers will have to provide edge-case workarounds until individual distributions fix their implementations.
The same goes for module managers. It's good to provide an open spec and encourage multiple frontends, but it's also good to provide a uniform interface for system administrators who just want to automate module installation for an application or framework. Python's easy_install wasn't ideal, but it provided exactly that sort of common denominator.
"Batteries included" makes it sound like I want my language to come with a playground, which I think is an unfair way to characterize the expectation that Perl 6 provide feature parity with the languages it competes with (i.e., Ruby 2 and Python 3).