[Maypole] Maypole startup sequence

[Dave Howorth]

Simon Flack wrote:
> This is very useful. But I want to reiterate what's in the first 
> paragraph and ask that you don't couple your applications to that 
> sequence. It is undocumented for a reason :)

I agree, applications shouldn't depend on anything that isn't documented 
elsewhere. I'll add something to that effect to my next draft.

But I think it is important to explain this stuff, particularly for 
debugging. When you have an application that won't start, or you're 
trying to extend some feature, it can be difficult to figure out what 
call you've put in the wrong place. And up until now, Maypole itself is 
not so completely robust that you can afford to treat it as a black box.

> In terms of backwards-compatability, I indend to continue supporting
> the request process outlines in Maypole::Manual::Workflow. But I
> might add some extra hooks to reduce the need to subclass. This'll
> probably happen after the upcoming release.

Sounds good.

Cheers, Dave