Larry writes an apocalypse, and Damian then sets about interpreting that to produce some 'real' code. And in doing this he trips over possible problems with the Apocalypse so he goes back to Larry for clarifications and/or bugfixes, and so the design evolves.

What we have here is a case of two heads being better than one. I am utterly convinced that without Damian's input the Perl6 design process would be even slower than it already is. I'm also expecting it to accelerate. A/E 3&4 are fundamental to the workings of Perl 6.

As for the length of time that E4 took, Damian has said that the first draft exposed some flaws in A4 and those took a good long while to fix. It can't have helped either that Damian hasn't been well.