|
I would expect that a larger number of smaller files would perform much better. The amount of memory and CPU power needed to keep track of the metadata for a few hundred files is negligible. The amount of RAM and CPU overhead involved in rendering a large DOM tree is huge. So the smaller the HTML files are, the better I would expect the book to perform, assuming all else is equal.
If that isn't the case with some particular reader, I'd be curious why. It probably involves somebody accidentally implementing bubble sort or something equally atrocious. :-)
|