Dear Bernhard,
Thank you for reacting so gracefully to my attempt at lightening
the tone of the discussion. I certainly did not mean to make light of
your and other people's concerns, not to cast doubts on the seriouness
of the problem.
As it happens, I made a very similar suggestion to yours (in jest
at the time!) in a discussion I had towards the end of the previous
millenium with several high-powered structural biologists (i.e., at
the time, successful crystallographers!) at a conference. There was
already concern, as the first multi-step structure solution systems
such as Tom Terwilliger's SOLVE and our own SHARP were beginning to
spread, and looking at the even earlier spread of SHELX in small-
molecule crystallography, that this would lead to a huge amount of
expertise and know-how in the field being diverted into software black
boxes and no longer circulating through human brains, creating hazards
of damaging misuse of these black boxes. Drawing a parallel with the
hazards of X-rays, I made the fairly obvious suggestion that access to
black-box software should be treated like access to synchrotrons and
should be granted only to users who would have successfully completed
a Safety Training test. This would have required that software
developers abide by the principle that any item of scientific
knowledge needed in order to "safely" use the know-how implemented in
their programs should be included in that test.
In practical terms this is what good documentation, and the good
use of documentation by users and their mentors, should consist of,
but there are quite a few obstacles in the way. The main one is the
"impatience factor" created by the competitive funding mechanism, be
it the funding of methods and software developers or that of the PIs
whose careers depend critically on results produced by such software.
The worst parental behaviour, as is well-known, it to "do things" for
children instead of letting them learn how to do them for themselves,
just because it saves time. In the case of "awsome PIs" not passing on
their crystallographic expertise to their team, it is even worse: they
entrust a third party (the software) to do those things for the junior
team members - so the software robs those junior members from learning
the safe use of it, and the PIs from learning how to communicate their
own expertise in that area. There are very laudible attempts at
plugging that gap through excellent Workshops (RapiData comes to mind,
but there are others), but my colleagues who take part as tutors in
such efforts invariably observe that the impatience of the PIs is
still tangible through that of many students to identify the software
that claims to require pushing the fewest buttons to do the job for
them.
The problems just brought up by the misfiring of MR that started
this thread are very much in the foreground of meeting organisers: see
for example the fowwling two sessions at the last ACA meeting:
http://www.amercrystalassn.org/2016-scientific-program#01.10
http://www.amercrystalassn.org/2016-scientific-program#01.12
The questions raised by the latter need some particularly careful
thought, especially the last sentence: the risk is not only that
technique development could be suffer from a reduction of fundamental
knowledge in the community (after all, modern software development and
distribution methods can amplify the work of just a few individuals to
meet the immediate needs), but that the general ability to use such
techniques safely could be compromised.
As things are at the moment, it is likely that the impatience
factor will continue to prevail, and that the software will keep
trying to become increasingly user-proof rather than the community
taking the trouble to make it more user-intelligible - because it
saves time now.
However - and this was the reason I thought of Alan Bennett's
quote yesterday - deploring this state of affairs is perhaps a
short-sighted longing for the way things were in some golden era. The
power cf computing is such that we have to learn how to use it better
rather than put limits on it, and things will never stay the same.
With best wishes,
Gerard.
--
On Sat, Sep 03, 2016 at 11:28:37AM -0700, [log in to unmask] wrote:
> Dear Gerard,
>
> thank you for your uplifting words. I admit being guilty of employing the
> rhetorical figure of hyperbole,
> while at the same time maintaining that too much metaphorical pee may stink
> up our shop.
>
> I entertain the vision of programs providing perhaps mandatory training
> (yes, test) for their use.
> Imagine being an unsupervised millennium generation, Facebook-YouTube-raised
> student, left to
> your own devices. Wouldn't you pick a program that offers required training
> on a video rather than
> one that lets you die stupid?
>
> E.g.: You want to build a ligand - pull down the menu - oops haven't viewed
> the prerequisite module test yet -
> do you want to proceed or probably build it wrong with the other programs.
> Your choice....
>
> If that proposition can be defended then maybe some effort spent on more
> than just providing optional
> (and only posteriorly read manuals) instead of hyperautomation can be
> justified in the course of program
> development. Might even have the benefit of saving the developers answering
> the same basic questions
> over and over, without having to resort to simply telling 'rtfm'. Which
> will become 'wtfv'...
>
> Cheers, BR
>
>
> -----Original Message-----
> From: Gerard Bricogne [mailto:[log in to unmask]]
> Sent: Saturday, September 03, 2016 10:58 AM
> To: [log in to unmask]
> Cc: [log in to unmask]
> Subject: Re: [ccp4bb] Another pifall of MR
>
> Dear Bernhard,
>
> As I sense from the tone of your message that you might be in danger of
> sinking into despondency, may I quote this little gem by the English
> playwright Alan Bennett:
>
> "I saw someone peeing in Jermym Street the other day. I thought, is
> this the end of civilization as we know it? Or is it simply someone
> peeing in Jermyn Street?"
>
> (as a bit of context, Jermyn Street used to be one of those emblematic
> streets in the centre of London, lined up with exclusive shops in which only
> members of the ruling class would ever be found, buying their top quality
> clothing - shirts, in particular, for that street).
>
> This case, too, may be shocking, but it it might not necessarily be
> "the end of civilization as we know it" ... :-) .
>
>
> With best wishes,
>
> Gerard.
>
> --
> On Sat, Sep 03, 2016 at 09:42:08AM -0700, Bernhard Rupp (Hofkristallrat
> a.D.) wrote:
> > > And in this way we can disguise a failure in proper training as a
> > > pitfall
> > of the automated MR computer programs
> >
> > Not to appear overly cynical -we have lost that training battle a long
> > time ago. Just look at some of the very basic questions asked on the
> > board and wonder - is there no supervisor providing any guidance? Some
> > of it may originate from that in certain environments it is not
> > customary to ask too many questions (particularly not those that might
> > challenge a superior).
> >
> > Then compare this with the postdoc job ads, which are best described
> > as looking for Jesus Christ - where are these geniuses supposed to
> > come from? This discrepancy between poor training, particularly
> > affecting structural biologist due to deficiencies ingrained already
> > early in the academic curricula, is quite striking.
> >
> > > re-program them to return a 'sorry, try again, no convincing
> > > solution
> > found' instead of returning the best solution
> >
> > I think this is a good idea. Require to sit through a lecture on
> > YouTube and pass a test to obtain a password for access to any expert
> > mode - at one's own risk. That way you instill some training even in a
> > hopeless case - any maybe it still works. Nothing to lose.
> >
> > Cheers, BR
> >
> >
> > Nicholas M. Glykos, Department of Molecular Biology
> > and Genetics, Democritus University of Thrace, University Campus,
> > Dragana, 68100 Alexandroupolis, Greece, Tel/Fax (office) +302551030620,
> > Ext.77620, Tel (lab) +302551030615, http://utopia.duth.gr/glykos/
--
===============================================================
* *
* Gerard Bricogne [log in to unmask] *
* *
* Global Phasing Ltd. *
* Sheraton House, Castle Park Tel: +44-(0)1223-353033 *
* Cambridge CB3 0AX, UK Fax: +44-(0)1223-366889 *
* *
===============================================================
|
