2003-03-28  Karl Hasselstrom  <kha@treskal.com>

	* report.tex.m4 (subsection{Divide-and-Conquer versus Newton
	Division}): Clarified captions of diagrams showing time for
	multiple divisions with same divisor.
	
	* report.tex.m4 (subsection*{Updated version}): Added a subsection
	to the preface where I explain that this is an updated version of
	the thesis.

	* report.tex.m4: Removed my e-mail address from the front page,
	since I just learned it will stop working a few months after I
	graduate. :-(

	* report.tex.m4: Added another 10 pt to height of \blankline,
	which determines distance between consecutive figures.

	* report.tex.m4 (section{Implementing Newton Inversion and
	Barrett's Algorithm}): Instead of claiming that GMP subroutines
	work on operands in-place, I now say "Since they take pointer
	arguments saying where to read input and write output, there is
	little overhead involved", which is closer to the truth.

	* report.tex.m4 (section{Test Runs}): Corrected test machine speed
	(was 1467 MHz, should be 1460 MHz).

	* report.tex.m4 (subsection{Divide-and-Conquer versus Newton
	Division}): Clarified what was meant by "but just one extra
	division ($k = 2$) was enough to bring it down from five to just
	below one million bits".

	* report.tex.m4 (section{Implementing Newton Inversion and
	Barrett's Algorithm}): Removed "extreme" from "where such extreme
	measures would be called for", since it was not really called for.

2003-03-09  Karl Hasselstrom  <kha@treskal.com>

	* report.tex.m4 (section{Karatsuba's Algorithm}): Fixed typo (a
	forgotten \beta^n was inserted).

	* report.tex.m4: Footnotes indicated with stars and crosses
	instead of numbers.