Earlier quoted context omitted.
> I don't think the RubyGems people were incompetent. They sat on a publicly disclosed vulnerability in the YAML parser for a week. The YAML parser itself was ridiculously designed to (essentially) eval() YAML. Those were the two active decisions of incompetence. On top of this, they built a massively central system that is widely trusted with no means of code verification whatsoever. There is no telling what people…
> On top of this, they built a massively central system that is widely trusted with no means of code verification whatsoever. https://github.com/rubygems/rubygems/blob/master/History.txt 0.8.11 / 2005-07-13: Added Paul Duncan's gem signing patch. They've had a mechanism for code signing for 8 years. Yes, they could require signing of all gems on the site, but the ability has been there for a long time.
As a counter-example, the maven central repository requires signatures, and caching Maven repositories validate those signatures.