[jsword-devel] XML Parsers

[DM Smith]

Please give me your thoughts:
Background:
    BibleDesktop uses Java WebStart as one of its "installers" and Java 
1.4 or higher.
    WebStart and Java 1.4 are written in such a way that the built-in 
xml parser cannot be replaced.
    Java 1.4 uses Crimson for its parser and one has to go to great 
lengths to replace it via a wrapper script.
    Java 1.5 (aka 5.0) uses Xerces for its parser and it is designed to 
be replaceable, even via WebStart.

Opportunity:
    Xerces has capabilities that Crimson does not. A few of these would 
be genuinely useful.
    For example:
         Xerces allows the programmatic specification of a schema. (This 
would allow us to have a debug mode that could validate fragments 
against a schema.)
         Xerces provides the ability to get line and offset information 
in error reports.

Do we go to Java 1.5? (At least for Java WebStart?)
Do we stick with source compatibility to Java 1.4?
Do we try to achieve source compatibility with Java 1.3? (This would 
allow us to target handhelds. We just need to replace regex and asserts.)
Do we use the latest and greatest Xerces? (It adds 1,176K + 190K to the 
install size.)