RE: Problem with font sizes on plots:

[Laurence Lurio]

Hi,

I think this was an answer to someone elses question?

Larry

-----Original Message-----
From: Ilan Boaz [mailto:address@hidden 
Sent: Monday, October 20, 2008 2:09 AM
To: address@hidden
Cc: Dmitri A. Sergatskov
Subject: Re: Problem with font sizes on plots:


Larry,

I think what Dmitry is trying to say is that if you expand the  
exponent as
exp(cx+d) = exp(cx)*exp(d)
then open the brackets (ax+b) you get
f(x) = a*x*exp(d)*exp(cx)*x + b*exp(d)*exp(cx) + e.
You can now redefine another set of constants as
a_new = a*exp(d)
b_new = b*exp(d)
Then
f_new(x) = a_new*x*exp(cx)*x + b_new*exp(cx) + e.
The problem of satisfying the 6 equations remains basically the same,
expect that the "names" of the constants have changed.

In other words, if you can solve the problem for f_new then you can  
solve if for the
original f(x) and vice versa. In other words, you really have only 4  
independent
constants: a_new, b_new, c, and e.

Now suppose you solved the problem for f_new and found a_new, b_new,  
c and e
-- there would be infinitely many ways to choose the original a, b,  
and d such that
a_new = a*exp(d)
b_new = b*exp(d)
For example, suppose a_new=b_new=1. Than one option is a=b=1 and d=0 and
another option is a=b=2 and d=ln(1/2), etc.

The fact that you used a software to obtain a, b, ... f that gave a  
function f(x)
that more-or-less satisfies the 6 equations does not imply that there  
is a a set of
constants that satisfies those 6 equations *exactly*. In fact, I  
suspect that in this
case the answer is that it is impossible to satisfy those 6 equations  
exactly.
Here's why: you really have only 4 independent constants but 6  
conditions.
It's like asking for a linear function y=ax+b to pass through 4  
points -- that can
happen only if those 4 points reside on a line, which is like a  
"miracle", i.e.,
it does not occur if you choose 4 points in an arbitrary way.

For 6 arbitrary-looking equations you would expect to have a function  
with 6
(independent) constants -- but in your case f(x) has only 4  
(independent) constants.
So I suspect there are no exact a, b, ... f that solve the problem.

Hope this helps.

Boaz


On Oct 14, 2008, at 9:07 AM, Dmitri A. Sergatskov wrote:

> On Tue, Oct 14, 2008 at 10:18 AM, Laurence Lurio <address@hidden>  
> wrote:
>> I would like to find a way to change the font size on the axis  
>> numbers in a
>> plot. I have tried using the Defaulttextfontsize command, e.g.
>> figure(1,'Defaulttextfontsize',18) but this only changes the font  
>> size on
>> the labels for the axes, not on the axis numbers.  In principle, I  
>> could
>> also try to work around this problem by printing the plot using the
>> -Sxsize,ysize option to make the plot size smaller and the axis  
>> numbers
>> relatively larger, but this also doesn't work.  Whenever I try to  
>> use the
>> print command I get an error "gdImageStringFT: Could not find/open  
>> font
>> while printing string".
>>
>
> Larry,
>
> There are few unrelated problems here. The easiest (IMHO)  
> workaround is
> to print to postscript and then convert to png (if you have to), e.g.:
>
> print ("file.eps", "-depsc2", "-FHelvetica:18")
>
> should work
>
>>
>>
>> Larry
>>
>
> Dmitri.
> --
>
> _______________________________________________
> Help-octave mailing list
> address@hidden
> https://www-old.cae.wisc.edu/mailman/listinfo/help-octave