[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ddlm-group] Relationship asmong CIF2, STAR,CIF1 and Python. . . .. .

Dear Colleagues

In reading through the overnight round of discussions, I am struck
by the potential for confusion that exists in our simultaneous
development of three inter-related frameworks - CIF, DDL and dREL,
and I suspect it's possible that I'm not the only one who has
occasionally been misled by transposing ideas on opposite sides of hte
boundaries between these frameworks (as well as that of the elusive
STAR - I'll document my particular blunder there in a separate posting
for the record).

My greatest regret about the manner in which we're attempting this
development is the lack of prototyping software to which we all have
access and can use as a test-bed to experiment with and evaluate the
effects of different approaches to the syntactic features. But I'm
aware of the circularity we're locked in - one can't design the
software without a reasonable complete starting specification...

Syd and Nick's proof-of-concept experiments, and James's subsequent
work with Nick have demonstrated, adequately in my opinion, that a
workable system is achievable, building on the existing STAR/CIF
formalism with modifications that are incremental rather than
revolutionary. It's on that basis that we have been trying to
anticipate and implement the remaining changes that should accommodate
a complete system working in that framework.

I think we have made terrific progress towards that goal: describing a
complete set of frameworks - CIF, DDL, dREL - built in the spirit of
the STAR formalism, that can achieve what we expect of them. We
probably haven't got everything right - only practical experience
can demonstrate that; but it's not for want of trying. I would
suggest that our energies in the next few months are best directed
towards accepting the currently agreed features, wrapping up the
discussion on the few remaining issues, expressing what we have as
reference documentation, and building initial applications to see if
there are unforeseen inconsistencies or contradictions in what we've
specified.

I have no doubt that some of the decisions we have reached will be
"wrong" with future hindsight, but I'm sure that will be true of any
set of decisions we reach based on any criteria. (Criterion 1: don't
ever listen to me. I always vote for the losing candidate in any
election. And I ALWAYS pick the wrong supermarket check-out line.) 

Those decisions that turn out to give rise to logical inconsistencies
will have to be revised in the light of experience; those that were just
"sub-optimal" may be usefully revised at the same time, or may just
have to be lived with as imperfectiosn.

I'm very heartened by James's confidence in the integrity of the
existing proposals, in the light of his dREL programming experience
(and I'm also struck by his conservatism to adopt Python formalisms
wholesale, given his practical experience in working with Python-based
CIF libraries and applications). I find a lot of what John says in
support of the current approach equally reassuring, both in terms of
his careful point-by-point analysis and his support for the usefulness
of developing the new system in the spirit of the original CIF: an
archive standard not susceptible to rapidly changing formulations.

I do appreciate many of Herbert's concerns based on practical
programming issues, and it is frustrating to be hobbled by syntactic
decisions that complicate programming (or, worse, introduce runtime
inefficiencies). But I come back to the point that we'll run into some
such problems somewhere, whatever we decide, and I'd rather we made
the "wrong" decision than starve like Buridan's ass, forever torn
between the equidistant heaps of straw and carrots.

Best wishes
Brian


On Sat, Jan 15, 2011 at 04:03:23PM +1100, James Hester wrote:
> dREL strings and elides have nothing to do with CIF2 syntax.  They are
> dREL syntax.  A dREL method is just an ordinary string for the CIF2
> parser, and has no special status.
> 
> On Sat, Jan 15, 2011 at 2:35 PM, Herbert J. Bernstein
> <yaya@bernstein-plus-sons.com> wrote:
>> Dear James,
>>
>>  Please reread the dREL string quoting specifications.
>> Please reread the dREL conventions on elides.  Let's
>> settle whether and in what way those items should
>> remain in the language.
>>
>>  Regards,
>>    Herbert
>>
>> =====================================================
>>  Herbert J. Bernstein, Professor of Computer Science
>>   Dowling College, Kramer Science Center, KSC 121
>>        Idle Hour Blvd, Oakdale, NY, 11769
>>
>>                 +1-631-244-3035
>>                 yaya@dowling.edu
>> =====================================================
>>
>> On Sat, 15 Jan 2011, James Hester wrote:
>>
>>> I welcome such analysis, which I believe is likely to lead to no more
>>> than a list of useful editorial changes to DDLm related to small
>>> changes in syntax since the DDLm draft document was produced.
>>>
>>> It is not trivial to produce a CIF2 version that is fundamentally
>>> unfriendly to DDLm and dREL, given that DDLm and dREL operate at a
>>> semantic level and we are producing a syntax document.  Only if our
>>> CIF2 syntax were to produce a different infoset (abstract
>>> datastructure) could some of the concepts in DDLm and dREL become
>>> poorly specified.
>>>
>>> On Sat, Jan 15, 2011 at 11:02 AM, Bollinger, John C
>>> <John.Bollinger@stjude.org> wrote:
>>>>
>>>> On Friday, January 14, 2011 2:30 PM, Herbert J. Bernstein wrote:
>>>>
>>>>>   I would suggest recalling that this CIF2 exercise began with
>>>>> tryin to define a DDLm/dREL-friendly version of CIF.  To the best
>>>>> of my knowledge, the current DDLm spec is at what is published
>>>>> on the IUCr web site under http://www.iucr.org/resources/cif/ddl/ddlm
>>>>> with draft documents from August 2008.  I propose that we all
>>>>> take the time to review those documents and make a conscious
>>>>> effort to decide if we wish to make use of them as a
>>>>> base for CIF2 or not.  They contain many points relevant
>>>>> to the current discussion.
>>>>
>>>> This is an eminently reasonable suggestion, and I have endeavored to take
>>>> it.  I find I need more time to formulate the results of my analysis, but
>>>> they do not lead me to the same conclusion that Herbert has reached, that we
>>>> should start over from scratch.  However, we can, perhaps, consider each of
>>>> the changes in our current draft, as Herbert prodded us to do, and evaluate
>>>> them against this group's mandate and objectives.  We might now draw
>>>> different conclusions about some of them.
>>>>
>>>> I decline to work on this over the weekend, and Monday is a U.S. holiday,
>>>> so it will be at least Tuesday before a I respond more fully.
>>>>
>>>>
>>>> Regards,
>>>>
>>>> John
>>>> --
>>>> John C. Bollinger, Ph.D.
>>>> Department of Structural Biology
>>>> St. Jude Children's Research Hospital
>>>>
>>>> Email Disclaimer:  www.stjude.org/emaildisclaimer
> -- 
> T +61 (02) 9717 9907
> F +61 (02) 9717 3145
> M +61 (04) 0249 4148
> _______________________________________________
> ddlm-group mailing list
> ddlm-group@iucr.org
> http://scripts.iucr.org/mailman/listinfo/ddlm-group
_______________________________________________
ddlm-group mailing list
ddlm-group@iucr.org
http://scripts.iucr.org/mailman/listinfo/ddlm-group


Reply to: [list | sender only]