This article shows how much progress we've made in languages and tools over 13 years. It's vastly more difficult to read large, c++ codebases with tons of memor

This article shows how much progress we've made in languages and tools over 13 years. It's vastly more difficult to read large, c++ codebases with tons of memory allocation, pointer arithmetic, macros, and complex locking schemes than much of the Java/Ruby/other mainstream language code written today. Also, it strikes me that this was written in an era where most large-scale systems were giant monolithic codebases sharing an address space, rather than distributed groups of services communicating over a network. Distributed systems are inherently easier to rewrite because a lot of the hard problems around abstraction, state, representation, etc. already have to have been solved by the architecture -- you can't exactly pass a naked void* over the wire.