RE: [xri] Very rough algorithm outline

[Sakimura Nat <n-sakimura@nri.co.jp>]

Got it.

=nat

________________________________________
差出人: Scott Cantor [cantor.2@osu.edu]
送信日時: 2009年6月18日 12:23
宛先: Sakimura Nat; 'XRI TC'
件名: RE: [xri] Very rough algorithm outline

Nat Sakimura wrote on 2009-06-17:
> But then, would it not mean that we start copy & pasting the spec?

No, it means you might have to add a non-normative part that explains why
the reference to the full spec doesn't mean you have to implement it.

All of the implementations out there know how to optimize for subtrees, and
they do so, but only with a lot of extra logic welded on to the complex
parts. If you don't care about the complex parts, the optimization is all
that's left.

The spec is still the spec. The trick is in understanding how to spot where
it's assuming generality that doesn't apply. With respect to excl c14n, that
subsection I keep pointing to is basically the optimized algorithm.

-- Scott