[Dev] DRAFT: Python Schema API proposal

[Phillip J. Eby]

At 03:21 PM 4/15/05 -0700, John Anderson wrote:
>I've probably written as much parcel XML as anyone, and don't particularly 
>like it, however, the vast majority of the parcel xml I've written turns 
>out to be creating instances, not creating schema, so I'm much more 
>concerned about how you intend to improve creating instances that populate 
>the repository during parcel loading when Chandler isn't running. I almost 
>never create new items programatically after the repository is populated, 
>instead I just make copies of existing items in the repository, so I also 
>don't care too much about the syntax of Item construction when Chandler is 
>running.

As I understand my charter for 0.6 from Katie and Lisa, I'm to focus on 
schema definition first.  I think that they'd like to see more of Alec's 
ideas for a CPIA-specific XML dialect before moving away from XML 
altogether for item instances.  But of course I'm sure they can speak more 
to that point than I can.

As far as I understand from Andi, however, none of the features he's done 
are restricted to schema items, and in fact if I understand his new 
Citem.COPYEXPORT correctly it might be able to help with the trunk copying 
stuff in the future as well.