[
Lists Home |
Date Index |
Thread Index
]
All good reasons (as is the suggestion from Eddie Robertson).
My next question was going to be whether the schema information needs to
be integrated with the instance (as the PSVI does) or if it can be held
separately and referenced as necessary. To me, this was an issue of both
memory efficiency and implementation complexity. However, the more I
thought about it, the more I realized this is purely an implementation
detail.
In a more traditional example, consider a SQL result set. In
metadata-based processing of a result set, you ask the result set what
the data type, name, etc. of the current column is. You have no way of
knowing whether that metadata is repeated for each column in each row
(tightly integrated with the data) or a single copy of the metadata is
held by the container (result set).
The same holds true for the PSVI. Assuming PSVI methods were added to
(for example) the DOM, the application doesn't know or care where the
PSVI information is held. While some of it is likely to be stored by
each node (e.g. validation information), others is likely to be stored
by the tree as a whole and referenced from individual nodes (e.g. type
information).
I think this is an issue during serialization, but (a) there seems to be
no reason to ever serialize the PSVI, and (b) it is easily serializable
in the form of the original schema and regenerated later.
(How did I ever get to sound like a fan of the PSVI? Or maybe I loved it
all along and just hated the name ... ;)
-- Ron
Jeni Tennison wrote:
>
> I can think of three ways in which I'd like to use information
> available in the PSVI within XSLT:
>
> [examples snipped]
|
