Re: [ccp4bb]: ESD of distances!

[lothar esser <lesser@helix.nih.gov>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***


> > > Why not?  You apparently had enough memory to do full-matrix in shelx,
> > > so you must therefore have ample memory for the PARST arrays.
> >
> >   My guess is he used CGLS as recommended for large molecules and not
> >   full-matrix least-squares.
>
> I had snipped the part where he said that he had used shelx to calculate
> esds.  That option only exists in L.S. mode, not in CGLS.
>
> It is possible to use L.S. refinement in block mode (BLOC) rather than
> true full-matrix, and ask for esds.   May I recommend to people that you

  Okay, it has been too long since I used CGLS for structure refinement
and it is clear that the recommendation to use CGLS is valid if the ratio
between observations and parameters is somewhat less than ~8. If it is
better one should use L.S. for all the good reasons you mentioned and not
even consider BLOC'ed least squares if it can be avoided.

And to support Rams in his quest: I'd be interested to know a solution to
his question too. ( Even if it means to modify PARST to perform dynamic
memory allocation - has anybody done it ? It is Fortran so it is possible
but not necessarily easy ).


Lothar