30-Dec-99 19:25 Klaus Weide wrote:
> On Fri, 31 Dec 1999, Leonid Pauzner wrote:
>> 30-Dec-99 10:23 Bela Lubkin wrote:
> 2) Only part of that memory is for the document constructed from the LIST
> temporary file. Apart other stuff, there are other documents in memory,
> at least the document from which the LIST was constructed - which would
> be similarly huge.
Ugh, right.
> Well, Bela gave us two approximate data points, so we can plot a straight
> line to get the (memory bytes / file bytes) increase:
>>> finish. Silly me. Process is up to 51MB in size and has only read
>>> about half of the 14MB file ... looks like it's actually going to push
> (74 - 51) / (12 - 7) = 4.6
> but the input numbers are only very approximate.
But overhead really large anyway.
Do we store highlighted text duplicated?
Do we store a list of URLs in HText, rather than a list of pointers?
