Re: ACTION-638: high level reorg suggestions

On 12/6/10 6:32 PM, "Pratik Datta" <pratik.datta@oracle.com> wrote:
>Since we are in the topic, one problem I have always had with the overall
>flow of the document, is that it requires you to understand the
>"Compatibility Mode" before you can understand the "2.0 mode".   But we
>are essentially deprecating the "Compatibilty mode".  So rather it should
>be the other way around  - i.e. most of the doc should be about 2.0 mode,
> and towards the end of the doc we can explain that whatever is in 1.0 is
>also supported, maybe we can have a separate top level section for that.

I think we have to decide whether to try to require people to read both
specs or not, because there's obviously a ton of duplication (e.g. all of
KeyInfo). If not, then we can't short-shrift the old material, but we can
reorder things to de-emphasize the old stuff.

>Right now most of the 2.0 stuff is inside "Section 6. Algorithms."
>embedded inside the 2.0 Transform.

That's one reason I'd like to pull that out to new sections inside section
6.

>  I would rather have the doc say -don't worry about meaning of
>Transforms, that is a just a hack we are doing to preserve backward
>compatibility, here is the new reference processing model
>Selection->Canonicalization->Digesting. And then later on we can describe
>what Transforms were and why we have deprecated them.

I think Frederick's suggestions are in that direction by splitting the
processing model discussions at the top. I didn't want to be that radical
with this first set of changes.

>(Not to start a gigantic bikeshedding discussion but, should this be
>called something other than "2.0 Mode", like "Streaming Mode"?)

I don't like the name (it was originally just a placeholder), but
Streaming Mode doesn't work, since that's not a requirement.

-- Scott

Received on Tuesday, 7 December 2010 00:06:50 UTC