I find it somewhat odd that Peter ascribes the valuing of stability and backwards compat to some weird Google-specific mindset, unheard of in other circles.
Instead, you can observe the same thing with many other ecosystems/platforms once they have reached a certain scale.
Ask Microsoft's Windows team or the Linux kernel maintainers about the importance of keeping existing programs running. Or maybe the stewards of other programming languages like C or C++ can tell you about the difficulty of evolving a language when you have a gazillion lines of pre-existing code to consider. How long does it take the world to phase out support for obsolete SSL/TLS versions, even in the presence of well-known weaknesses?
It seems to me that the more users something has, the more thought has to be put into how not to break them.
Isn't programming in the large exactly the kind of thing Go set out to tackle? Placing a high value on stability appears only logical then.
Instead, you can observe the same thing with many other ecosystems/platforms once they have reached a certain scale. Ask Microsoft's Windows team or the Linux kernel maintainers about the importance of keeping existing programs running. Or maybe the stewards of other programming languages like C or C++ can tell you about the difficulty of evolving a language when you have a gazillion lines of pre-existing code to consider. How long does it take the world to phase out support for obsolete SSL/TLS versions, even in the presence of well-known weaknesses?
It seems to me that the more users something has, the more thought has to be put into how not to break them.
Isn't programming in the large exactly the kind of thing Go set out to tackle? Placing a high value on stability appears only logical then.