[Jmol-developers] Re: detecting backbone atoms

[timothy d. <mol...@ma...>]

at 2:42 pm EDT on (Wednesday) 29 October 2003 Dan Bolser said:

> Miguel said:
> >> I found rasmol *very* slow at selecting multiple sets of atoms,
> > I honestly have no idea how the speed of the Jmol interpreter compares
> > with the RasMol interpreter.
> >
> > By saying 'selecting multiple sets of atoms' I assume that you are
> > referring to atom expressions with OR and AND operations. I don't
> > understand how RasMol would be slow at that.
> 
> It is a problem of defining large sets of atoms, which needs to
> be broken up into multiple define commands
> 
> 
> I.e.
> 
> define x something_to_select_which_fills_the_line_up
> 
> define x x, teh_rest_of_the_select_over_multiple_lines
> 
> This causes an n**2 comparison of 'defined flag' searches,
> which is a real pain for big defines.
> 

but the slowness is not limited to define commands that span multiple lines,
as Dan describes above.  for example, a 'simple' select command like 

select model=1 and protein

can be quite slow in RasMol & Chime, if there is a significant number of
atoms.

'simple' at the user level, that is.  at the source code level, it may be very
complex.

regards,

:tim

-- 
timothy driscoll
molvisions - molecular graphics & visualization
<http://www.molvisions.com/>
usa:north carolina:wake forest