[sdpd] Re: A comment on indexing

[Robin Shirley <R.Shirley...@surrey.ac.uk>]

What Armel actually said was that indexing is not *the* main problem but 
remains *one of* the main problems.  I'd go along with that.

Similarly I agree with Ken's reasoning, though not completely with his
emphasis.  Indexing isn't *the* (i.e. only) big problem.  Similarly,
structure solution from powders isn't yet a done deal, though the field is
thriving.

Probably it's too much to expect that structure solution ever *will* be "a
done deal", bearing in mind that pathological structures are quite common,
and that not all ofthem can be solved with single-crystal data either.

If we are talking about synchrotron data from well-crystallised single solid
phases, without significant unidentified impurities, especially if not in the
lowest symmetry (I'm just saying *if* here), then indexing shouldn't very
often be a problem (though sometimes it still will be).

For many (most?) people, though, life isn't like that.  With laboratory
data, with a limited number of broadened lines, or where one may have other
unknown phases (often hard to avoid with polymorphs, natural specimens or the
products of solid-state reactions) then it can be quite another story.

Though it has far fewer unknown parameters, I'm not sure that indexing is
actually an easier problem than structure solution, and it's certainly had
far fewer resources directed at it.  I'd like to see the last aspect change,
but in the first instance that will probably require some changes of
priorities within the powder diffraction community.

As Armel says of the indexing problem: "its peculiarity is that it is
occurring first - no solution at this step and you cannot pursue your
quest".

Robin Shirley

-------------------------------------------------------------

> From:          K.Shankland...@rl.ac.uk
> To:            sdpd...@egroups.com
> Date:          Sun, 6 Feb 2000 15:25:46 -0000
> Reply-to:      sdpd...@egroups.com
> Subject:       [sdpd] A comment on indexing

> Lachlan said
> 
> >If the main(?) problem people encounter with solving from
> >powders is reliable powder indexing - what percentage will
> >ever get through to trying solve using programs such as DASH.
> >(the above Round Robin was on a phase where the cell was given -
> >as it was a "solving" round robin - not an indexing round robin)
> 
> I agree with Armel - who says that indexing is the big problem ? 
> Lets keep things in perspective !  Structure solution is far 
> from a "done deal". Yes, global optimisation and enhanced 
> direct methods have transformed what might be considered to be 
> tractable by powder diffraction, but there is still plenty 
> of work to be done at this stage. 
> 
> Of the many compounds that we have solved over recent years, 
> the vast majority of them have indexed first time, albeit
> with mostly synchrotron data.  I suspect that there are just
> as many indexed patterns that have not been solved yet as there
> are unindexed patterns.
> 
> However, what is certainly true to say is that the ability 
> to tackle and solve much larger structures invevitably leads 
> us into an area where the unit cells are much larger and
> existing methods are perhaps not so effective.  So it
> seems likely that increased failure rates in indexing will
> be a consequence of the success of global optimisation 
> (and other) methods in structure solution.
> 
> Best,
> 
> Kenneth
>