Hi,
Setting parameter limits equal in a TF1 seems to cause a fit to logically
fail. pol1::p0 will always become the number, 288.074, for the profile
histogram, fixpar.C, no matter what fixing value I choose. Setting
ParLimits tolerably apart works, however.
hand built root-3.5.3 on:
redhat-8.0
gcc (GCC) 3.2 20020903 (Red Hat Linux 8.0 3.2-7)
AMD Athlon(tm) Processor 1994.691 MHz 256 KB cache
running root on a terminal with:
8.0 Mb memory
redhat-6.1 ctwm /*no desktop, only the window manager*/
vt100 xterm
bash-2.05b$ root -b -q fixpar.C fixpar.cpp
...
root [0]
Processing fixpar.C...
Processing scripts/fixpar.cpp...
Note: File "iostream" already loaded
288.074
288.074
1.1
/*fixpar.cpp*/
{
#include <iostream>
using namespace std;
htemp->Fit( "pol1", "q" );
pol1->FixParameter( 0, 1.0 );
htemp->Fit( "pol1", "q" );
cout << pol1->GetParameter( 0 ) << endl;
pol1->SetParLimits( 0, 1.0, 1.0 );
htemp->Fit( "pol1", "q" );
cout << pol1->GetParameter( 0 ) << endl;
pol1->SetParLimits( 0, 0.9, 1.1 );
htemp->Fit( "pol1", "q" );
cout << pol1->GetParameter( 0 ) << endl;
}
Justin
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:10 MET