At 11:35 AM 1/5/00 -0700, [EMAIL PROTECTED] wrote: >1) DOM will be bigger on platforms with a 32 bit wchar_t
Significantly bigger? Often not. If you look at a typical DOM implementation, when you load in a doc of size X, you typically burn N*X memory, where N ranges from 3 to 20. Maybe higher in perl :) Since a DOM implementation probably only stores each character once (we hope), the overhead is mostly tree-pointer bookkeeping. The conclusion is that the character-data text itself is not obviously a dominating factor in DOM size. Now obviously, this varies with the structure of the doc. If you have <start>.... 500k of text ...</start> then the character size probably dominates. But in general, the statement "DOM will be bigger" is misleading enough not to be useful. The rest of Dean's comments make sense. And I'm horrified that at this late date HP is storing anything but Unicode in wchar_t's. Sigh. -Tim
