[jdom-interest] Very Large Documents

[Tim & Mary Weatherstone]

Hello,
Is there any interest and/or expertise in handling v large documents, 
100,000+ elements and 40M+ ? I have attempted a first-cut of a solution 
by indexing the first rung of children, and then allowing selection by 
this index. In so doing I have reduced the size effectively to about 5% 
of the original. A bit of a 'noddy' solution perhaps ? or a sensible 
way to go about it ? I have had to increase the heap size to approx 
300M to achieve this.
Thanks,
Regards,
timW