Re: LaTeX suddently blows up .eps images???

Donald Arseneau <asnd@xxxxxxxxx> writes:

On Nov 27, 2:08 am, David Kastrup <dak@xxxxxxx> wrote:
Donald Arseneau <asnd@xxxxxxxxx> writes:
On Nov 25, 1:01 pm, Dan <lueck...@xxxxxxxx> wrote:
On Nov 25, 12:11 pm, "I.N. Galidakis" <morph...@xxxxxxxxxxxx> wrote:
100/75 is 1.25,

I thought TeX was supposed to be *good* at math.

Typesetting it.

But try this:

David, as someone mentioned recently, your humor quotient is at an
ebb.

\dimen1=0.01\dimen0

You fall into a trap thinking 0.01 is an exact number.

Do you really think I don't know how this result comes about? Hint: I
am the one who reported problematic values (and proposed the
corrections) for the nc and nd units available in PDFTeX, and in the
course of doing that, corrected Knuth's wrong documentation of the
requirements for numerator and denominator for those ratios.

On a computer it is not.

It certainly is. For _TeX_ (which _here_ chooses to use 15.16 bit
fractionals, whereas "in" is an exact ratio, to be truncated after given
a multiplier) it is not.

So _TeX_ falls into the trap thinking 0.01 is an exact number, namely
0.0099945068359375. Which makes it bad at math.

\dimen2=1in

Yes, that (non-) rounding is silly.

100in are 7227pt. A hundredth of that is 72.2303pt.

A hundredth cannot be calculated.

Try dividing by 100. It is not that hard. If you do it rounding to the
nearest 2^{-16}, you'll arrive at the value properly called 72.27pt by
TeX. Whereas 1in (which happens to be exactly 72.27pt) is called
something different by TeX.

But at least the round off is not so bad as to think 100/75 is 1.25.

I have had quite bad cases of rounding...

--
David Kastrup, Kriemhildstr. 15, 44793 Bochum
UKTUG FAQ: <URL:http://www.tex.ac.uk/cgi-bin/texfaq2html>
.

Relevant Pages
Loading