Discussion List Archives

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

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

Dear Colleagues,

   Just to be clear -- I think James did excellent work
on his DDLm/dREL software.  That is why I use it in
CBFlib.  However, that history is one of support for
what are essentially CIF1 files.  Somehow we have
now gotten into something very different from CIF1,
breaking the promise on the IUCr web site, whence
my desire to ensure that we are in agreement on
our goals.  I am not questioning what James has done,
just whether what we have been doing recently is
consistent with that very fine initial effort.
I am not proposing to go back to 1991, or even 2005,
but to 2008, and to re-examine what we have done
since then.

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 Tue, 18 Jan 2011, Herbert J. Bernstein wrote:

> Once we settle our goals, I will be happy to explore
> how well value generation does or does not work
> until the resulting scheme for CIF1 documents.  Right
> now I see a conflict when you cannot write the alias
> so that the dictionary tag is the CIF1 tag rather
> than the CIF2 tag, but if there turns out to be a
> way to conform both to CIF2 and to work fully with
> value generation in a CIF1 context, that would
> achieve the goal of my suggestion.   First it
> would help to know if full CIF1 interoperability
> is or is not one of our goals.
>
> =====================================================
> 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 Tue, 18 Jan 2011, James Hester wrote:
>
>> David's approach is perfectly workable for the tasks Herbert
>> describes, and is the proper way to operate in the CIF framework.  We
>> are agreed that the alias mechanism operates adequately for validating
>> existing values in a CIF1 file using a DDLm dictionary.  Herbert's
>> concern is that the DDLm framework cannot produce values for missing
>> CIF1 tags.  This is not true.  Given a missing CIF1 tag, the
>> validation software looks for a dataname for which that tag is given
>> as an alias in the DDLm dictionary, and then calculates the value of
>> the DDLm dataname thus found.  There is absolutely no need for any new
>> command on the dREL level.  All that is required is for DDLm
>> dictionaries to include equivalent tags, which is not a software
>> issue.
>> 
>> The CIF2-DDLm-dREL system works as advertised.  No changes are indeed
>> required to archival datafiles in order to apply DDLm dictionaries. It
>> is in fact transparently simple to apply DDLm dictionaries to CIF1
>> data files.  There is certainly no justification for revisiting CIF2,
>> or fiddling around with dREL.
>> 
>> Work commitments prevent me from answering in a more timely fashion
>> the various incorrect statements and interpretations that have been
>> flying around over the weekend.  I will be attempting to do this over
>> the next few days, and also to guide these discussions to some
>> conclusion.
>> 
>> On Tue, Jan 18, 2011 at 11:14 AM, Herbert J. Bernstein
>> <yaya@bernstein-plus-sons.com> wrote:
>>> Dear David,
>>> 
>>>  The promise on te IUCr website is very clear and in bold face
>>> 
>>> "No changes are required in existing archival data files in order to apply
>>> domain dictionaries written in DDLm"
>>> 
>>> The change I have proposed or something of similar power is required in
>>> order to be able apply DDLm dictionaries fully to CIF1 files so that both
>>> validation and filling in or missing values can be done in terms of the
>>> actual tags in the files, not just in terms of tags acceptable under CIF2.
>>> 
>>> I was willing to accept the sacrifice of such full functionality as 
>>> promised
>>> when we did not have a full technical solution on the table. Now we have
>>> one. If somebody has a better technical solution, that is wonderful, but I
>>> cannot and will not support retreating from that promise again, not that 
>>> we
>>> seem to have a way to keep it.
>>> 
>>> I do not understand what you mean by DDLm not being forwardly compatible,
>>> but I do under stand what applying a DDLm dictionary to a file means, and 
>>> if
>>> that file is a CIF1 data file, then we need a way to have the DDLm
>>> dictionary, which is itself a CIF2 file, work with the CIF1 tags directly 
>>> in
>>> its methods.  If all the methods can do is work in terms of DDL2 data 
>>> names
>>> then they cannot populate missing CIF1 values and their method-generated
>>> validation error messages (as opposed to control program generated
>>> validation error messages) cannot be done in terms of CIF1 tags, making 
>>> them
>>> much less useful.
>>> 
>>> I don't see how to use aliases to do this, unless we extend the alias
>>> definition to essentially do the same thing as the change I am proposing 
>>> for
>>> dREL, and make the necessary changes in dREL to make use of that alias
>>> information.
>>> 
>>> 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 Mon, 17 Jan 2011, David Brown wrote:
>>> 
>>>> Herbert,
>>>> 
>>>> I had assumed that the DDLm would be backwardly compatible, but not
>>>> forwardly compatible.  The _alias mechanism allows a DDLm program to read
>>>> CIF1 and CIF2 files and store the information under DDLm names.  There 
>>>> may
>>>> be some problems with datanames that include reserved characters, but the
>>>> lack of a magic number at the head of the file will warn the program of
>>>> this
>>>> possibility.  Once the file is read, all the features of DDLm are
>>>> available.  The program can then output a CIF3 (or CIFm) file without
>>>> further problem.
>>>> 
>>>> Whatyou want is to have the DDLm-based program write out a CIF1 or CIF2,
>>>> according to where it found the input or according to some other
>>>> instruction.  This could be done again by consulting the _aliases, but of
>>>> course there may not be datanames for all the items it calculates.  By
>>>> outputing the information under the appropriate _alias dataname, it 
>>>> should
>>>> be possible to output a CIF1 or CIF2.  The categories should take care of
>>>> themeselves.  The only place there would be difficulty is in the loops,
>>>> but
>>>> the DDLm rules aer stronger than the CIF1 and CIF2 rules, so the loops
>>>> should not violate the less stringent earler standard.  In any case I am
>>>> not
>>>> sure how your import..as would work in this mode (I would have thought an
>>>> export..as would be more appropriate).  The problem is that in the
>>>> relevant
>>>> items there are always two, and sometimes as many as four, _aliases which
>>>> means that dREL will become extremely cumbersome and would have to know
>>>> which _import to apply.
>>>> 
>>>> But this is not really what we are talking about the moment.  If we want
>>>> compatiblity at the syntax level we should be worrying about the use of
>>>> reserved characters in the older datanames (which we have already
>>>> discussed), not how we should modify dREL.  It would be better to use
>>>> _aliases even if we have to provide information about their original
>>>> categories.  This is by far the cleanest place - when we are reading or
>>>> writing a CIF - not when we are in the middle of applying a method.
>>>> 
>>>> David
>>>> 
>>>> Herbert J. Bernstein wrote:
>>>>      Dear David,
>>>> 
>>>>        The alias mechanism in the DDLm document of 2008 does
>>>>      not seem to do the job of getting around the restrictions
>>>>      that were imposed for CIF2 on the CIF files themselves,
>>>>      invalidating basically any tag that does not conform
>>>>      to the necessities of identifer syntax in a dREL method.
>>>>      What I am proposing as a way for a dREL method itself
>>>>      to cope with an arbitrary CIF1 tag, allowing true
>>>>      CIF1 documents handled.  This is important to, for
>>>>      eaxmple, ensure that when a DDLm method is populating
>>>>      a category in a CIF1 document, that it populates the
>>>>      actual original CIF1 category and not the category
>>>>      with the CIF2 alias name.  This will allow the power
>>>>      of DDLm and dREL to be applied to documents that still
>>>>      need to be processed by existing software that is
>>>>      not yet aware of new CIF2 names for various categories
>>>>      and tags, and therefore should not automatically migrate
>>>>      to the alias name.  The current form of alias seems
>>>>      to be purely a one-way path.
>>>> 
>>>>        Is there a later extension of alias after the 2008
>>>>      document?
>>>> 
>>>>        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 Mon, 17 Jan 2011, David Brown wrote:
>>>> 
>>>>            This is a pssing comment on the first paragraph of
>>>>            Herbert's email to set
>>>>            the record straight.
>>>> 
>>>>            David Brown
>>>> 
>>>> 
>>>> 
>>>>            Herbert J. Bernstein wrote:
>>>>                    What I propose is that we add one new
>>>>            statement to dREL,
>>>>                  similar to the python import .. as ..., which
>>>>            allows a module in
>>>>                  python that actually exists under one name to
>>>>            be used in a chunk
>>>>                  of pythin code as if it had a different name.
>>>>            One possible
>>>>                  syntax could be to simply extent the python
>>>>            import statment
>>>>                  allowing elides in the identifier name, such
>>>>            as
>>>> 
>>>>                    import cell\[\] as CELL
>>>> 
>>>>                  or some other quoting mechanism to allow CIF1
>>>>            names that are not
>>>>                  valid internally in dREL to be handled by dREL
>>>>            algorithms.
>>>> 
>>>>            We discussed this many moons ago.  For the most part
>>>>            Herbert's suggestion is
>>>>            already accommodated by the _alias feature of DDLm.
>>>>            This allows any CIF1 or
>>>>            CIF2 dataname to be recognized as equivalent to a
>>>>            DDLm dataname at the timw
>>>>            of reading.  Thus a CIF1 or CIF2 item will be
>>>>            treated exactly the same way
>>>>            in the program as if it were in a file written using
>>>>            a DDLm dictionary.
>>>>            There may be one or two places where a DDLm driven
>>>>            program might have
>>>>            difficulty, but they will not be common.  None come
>>>>            immediately to mind.
>>>>            The main problem will be trying to read the many
>>>>            archived CIFs that are not
>>>>            well formed, but there should be no need for a major
>>>>            reformatting of the
>>>>            existing archive.
>>>> 
>>>> 
>>>>            ## End of my comments
>>>> 
>>>> 
>>>> 
>>>>                    This does not mean that a DDLm dictionary
>>>>            would somehow have
>>>>                  to contain the identifiers that have been
>>>>            banned under CIF2. The
>>>>                  dictionary could work entirely under CIF2
>>>>            rules.  It just means
>>>>                  that the data being manipulated under control
>>>>            of the dictionary
>>>>                  could include unmodified CIF1 data files and
>>>>            that the algorithms
>>>>                  in dREL could be applied to the entire range
>>>>            of CIF1 tags.
>>>> 
>>>>                    There is no import statement currently in
>>>>            dREL, so there is
>>>>                  not conflict with the existing spec.  Other
>>>>            than providing a
>>>>                  symbol table translation, there should be no
>>>>            impact of this
>>>>                  change on the structure and syntax of dREL.
>>>> 
>>>>                    While this is an othogonal feature, I do
>>>>            intend to also
>>>>                  propose that DDLm dictionaries we used to
>>>>            process CIF1 files
>>>>                  that include CIF1-style string, e.g. the
>>>>            famous 'O'', so that
>>>>                  the promise made for so many years on the IUCr
>>>>            web site of being
>>>>                  able to process archived CIF1 data files with
>>>>            DDLm can be kept.
>>>> 
>>>>                    I think CIF1 (both DDL1 and DDL2 CIF)
>>>>            coupled to DDLm could be
>>>>                  a very powerful and useful combination that
>>>>            could be introduced
>>>>                  with minimal disruption to existing software.
>>>>            I understand that
>>>>                  others may prefer a more drastic change. I
>>>>            hope they will see
>>>>                  the value in allowing this more cautious and
>>>>            conservative
>>>>                  approach to go forward under the CIF banner,
>>>>            even if they
>>>>                  themselves intend to pursue their own, more
>>>>            drastic set of
>>>>                  changes, also under the CIF banner.
>>>> 
>>>>                    If this is agreeable, I will undertake to
>>>>            have a demonstration
>>>>                  version of DDLm/dREL fully coupled to CIF1
>>>>            using the import ...
>>>>                  as ... statement to show and discuss with
>>>>            those who are
>>>>                  interested in Madrid.
>>>> 
>>>>                    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 Sun, 16 Jan 2011, Herbert J. Bernstein
>>>>            wrote:
>>>> 
>>>>                        Dear Simon,
>>>> 
>>>>                         If I read your message correctly, you
>>>>            seem to be
>>>>                        saying that the design objectives of
>>>>            CIF2 are:
>>>> 
>>>>                         1.  enchance CIF as a data source
>>>>                         2.  accomodate DDLm
>>>>                         3.  do 1 and 2 in a way that does not
>>>>            complicate
>>>>                        CIF as a data source
>>>> 
>>>>                         I don't understand what you mean by
>>>>            "enhance CIF as
>>>>                        a data source". This is the first time I
>>>>            recall
>>>>                        hearing that objective.  Please clarify.
>>>> 
>>>>                        In terms of present practice, I had
>>>>            thought that the
>>>>                        major current uses of CIF to be:
>>>> 
>>>>                         1.  As a language for publication of
>>>>            papers in IUCr
>>>>                        journals;
>>>>                         2.  As a language for submission of
>>>>            data to CCDC;
>>>>                         3.  As a language for submission of
>>>>            data to the
>>>>                        PDB;
>>>>                         4.  As a data harvest language for
>>>>            CCP4; and
>>>>                         5.  As a language for formatting image
>>>>            data from
>>>>                        Dectris detectors,
>>>>                        with some emerging use for data
>>>>            management of
>>>>                        synchrotron data
>>>>                         ( Please fill in uses I have missed
>>>>            ...)
>>>> 
>>>>                        which would seem to me to represent a
>>>>            large current
>>>>                        investment in data management software
>>>>            that depends
>>>>                        critically on stability and reliability
>>>>            of CIF
>>>>                        representation of both numeric data and
>>>>            text.  As I
>>>>                        understand it, DDLm and dREL can
>>>>            contribute to these
>>>>                        current uses by enhancing the reliabiity
>>>>            of
>>>>                        validation of data in these contexts,
>>>>            especially in
>>>>                        uses 1, 2 and 3, above. To me, that
>>>>            would  seem to
>>>>                        change the objectives to:
>>>> 
>>>>                         1.  Accomodate DDLm and dREL; and
>>>>                         2.  Do this in a way that keeps
>>>>            required changes to
>>>>                        existing
>>>>                        archives to a minimum; and
>>>>                         3.  Do this in a way that allows as
>>>>            much existing
>>>>                        CIF software
>>>>                        as possible to continue to operate
>>>>            reliably; and
>>>>                         4.  To the extent that changes will be
>>>>            needed in
>>>>                        archives and
>>>>                        software, provide a clearly understood
>>>>            mechanism for
>>>>                        making
>>>>                        those changes, with as much support
>>>>            software as
>>>>                        possible; and
>>>>                         5.  Subordinate to the above, add new
>>>>            features to
>>>>                        CIF that may
>>>>                        encourage broader use and more software
>>>>            support
>>>> 
>>>>                        Please tell me what I have misunderstood
>>>>            in this.
>>>> 
>>>>                        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 Sun, 16 Jan 2011, SIMON WESTRIP
>>>>            wrote:
>>>> 
>>>>                              I believe that the CIF2 syntax
>>>>            changes
>>>>                              enhance CIF as data source,
>>>>                              even with the restriction to the
>>>>                              contents of  ' and " delimited
>>>>            values,
>>>>                              which
>>>>                              I suspect will be the main source
>>>>            of
>>>>                              incompatability between archived
>>>>            CIFs
>>>>                              and CIF2. What we havent acheived
>>>>            is
>>>>                              100% compatability of CIF1 with
>>>>            CIF2,
>>>>                              but having agreed that CIF2 is
>>>>            distinct
>>>>                              from CIF1 and minimized the
>>>>                              incompatability,
>>>>                              I do not see the changes as
>>>>            disruptive
>>>>                              or retrograde.
>>>> 
>>>>                              DDLm methods will not be
>>>>            applicable to a
>>>>                              significant percentage of a CIF
>>>>                              data source
>>>>                              (those 'free text' fields used to
>>>>            report
>>>>                              experimental details etc.). I
>>>>                              believe other approaches
>>>>                              will be needed to enhance their
>>>>            content
>>>>                              (referencing other data items from
>>>>                              an item etc).
>>>>                              Unicode support in CIF2 is
>>>>            beneficial in
>>>>                              this area, and the new alternative
>>>>                              delimiters may
>>>>                              prove convenient. The list and
>>>>            table
>>>>                              structures may also be exploited.
>>>> 
>>>>                              So from my point of view, though I
>>>>                              understood the aim of CIF2 was to
>>>>                              accommodate DDLm,
>>>>                              it turns out that it does offer a
>>>>            little
>>>>                              bit more. I would still like the
>>>>                              problem of including all
>>>>            delimiters
>>>>                              in a value to be solved, and
>>>>                              line-folding to be included. On my
>>>>            'wish
>>>>                              list'
>>>>                              are value concatenation
>>>>                              and a value-referencing mechanism
>>>>            for
>>>>                              use with CIF as a data source.
>>>> 
>>>>                              I know this doesnt really answer
>>>>            your
>>>>                              question. What we are *not* trying
>>>>            to
>>>>                              do is complicate
>>>>                               CIF as a data source?
>>>> 
>>>>                              Cheers
>>>> 
>>>>                              Simon
>>>> 
>>>> 
>>>> ___________________________________________________________________________
>>>> 
>>>>                              _
>>>>                              From: Herbert J. Bernstein
>>>>                              <yaya@bernstein-plus-sons.com>
>>>>                              To: Group finalising DDLm and
>>>>            associated
>>>>                              dictionaries <ddlm-group@iucr.org>
>>>>                              Sent: Sunday, 16 January, 2011
>>>>            2:05:59
>>>>                              Subject: Re: [ddlm-group]
>>>>            Relationship
>>>>                              asmong CIF2, STAR, CIF1 and
>>>>            Python. .
>>>> 
>>>>                              Dear Simon,
>>>> 
>>>>                                I have reviewed the record on
>>>>            this,
>>>>                              and I have trouble finding
>>>>                              a strong basis for most of the
>>>>            decisions
>>>>                              made other than arguments
>>>>                              from the authority of established
>>>>            past
>>>>                              practice in STAR and dREL.
>>>>                              Now that those arguments have
>>>>            turned out
>>>>                              not to have been what they
>>>>                              seemed, I have trouble accepting
>>>>            the
>>>>                              resulting conclusions without
>>>>                              new, clearly stated arguments
>>>>            based in
>>>>                              the functionality we are
>>>>                              trying to achieve, but it is now
>>>>            not
>>>>                              even clear what functionality we
>>>>                              are
>>>>                              trying to achieve.
>>>> 
>>>>                                Perhaps you can help me -- what
>>>>            are we
>>>>                              trying to do in defining
>>>>                              CIF2?
>>>> 
>>>>                                Regards,
>>>>                                  Herbert
>>>> 
>>>> 
>>>>                              At 11:55 PM +0000 1/15/11, SIMON
>>>>            WESTRIP
>>>>                              wrote:
>>>>                              >That's a question I asked myself
>>>>            when I
>>>>                              first joined this discussion group
>>>>                              :-)
>>>>                              >
>>>>                              >To be honest Herbert, I do not
>>>>            feal
>>>>                              qualified to contest anything to
>>>>            do
>>>>                              with
>>>>                              >dREL, nor DDLm, given that I
>>>>            havent
>>>>                              worked on them and respecting the
>>>>                              >years of effort that went into
>>>>                              development thus far. The same is
>>>>                              >true with respect to some of
>>>>                              >the changes to CIF syntax that
>>>>                              invalidate CIF1: I have accepted
>>>>            that
>>>>                              >they are necessary
>>>>                              >to facilitate implementation of
>>>>            dREL
>>>>                              methods.
>>>>                              >
>>>>                              >Cheers
>>>>                              >
>>>>                              >Simon
>>>>                              >
>>>>                              >
>>>>                              >
>>>>                              >From: Herbert J. Bernstein
>>>>                              <yaya@bernstein-plus-sons.com>
>>>>                              >To: Group finalising DDLm and
>>>>                              associated dictionaries
>>>>                              <ddlm-group@iucr.org>
>>>>                              >Sent: Saturday, 15 January, 2011
>>>>                              23:14:36
>>>>                              >Subject: Re: [ddlm-group]
>>>>            Relationship
>>>>                              asmong CIF2, STAR, CIF1 and
>>>>            Python.
>>>>                              .
>>>>                              >
>>>>                              >Why not?  It is almost that right
>>>>            now.
>>>>                              >
>>>>                              >
>>>>                              >At 11:08 PM +0000 1/15/11, SIMON
>>>>                              WESTRIP wrote:
>>>>                              >>True - but I can't see dREL
>>>>            becoming
>>>>                              pyREL at this stage?
>>>>                              >>
>>>>                              >>
>>>>                              >>
>>>>                              >>From: Herbert J. Bernstein
>>>> 
>>>> 
>>>>  >><<mailto:yaya@bernstein-plus-sons.com>yaya@bernstein-plus-sons.com>
>>>>                              >>To: Group finalising DDLm and
>>>>                              associated dictionaries
>>>> 
>>>>            >><<mailto:ddlm-group@iucr.org>ddlm-group@iucr.org>
>>>>                              >>Sent: Saturday, 15 January, 2011
>>>>                              22:57:17
>>>>                              >>Subject: Re: [ddlm-group]
>>>>            Relationship
>>>>                              asmong CIF2, STAR, CIF1 and
>>>>            Python.
>>>>                              .
>>>>                              >>
>>>>                              >>Dear Simon,
>>>>                              >>
>>>>                              >>  But dREL already shares much
>>>>            of
>>>>                              Python syntax and data structures,
>>>>                              >>but, being significantly
>>>>            mutated,
>>>>                              lacks the software support and
>>>>                              >>documentation that Python has.
>>>>            Anyone
>>>>                              who has to work with the
>>>>                              >>methods in a DDLm dictionary
>>>>            would be
>>>>                              much better off if we
>>>>                              >>simply made Python work with
>>>>            DDLm.  We
>>>>                              would gain large libraries
>>>>                              >>of pre-written utilities, tools
>>>>            to
>>>>                              test code fragments interactively,
>>>>                              >>and a lot more time to do
>>>>            science or
>>>>                              whatever we are actually
>>>>                              >>funded to do.
>>>>                              >>
>>>>                              >>  Regards,
>>>>                              >>    Herbert
>>>>                              >>
>>>>                              >>
>>>>                              >>
>>>>                              >>At 10:35 PM +0000 1/15/11, SIMON
>>>>                              WESTRIP wrote:
>>>>                              >>>As far as I can see, parsing
>>>>            DDLm
>>>>                              into an object stucture is fairly
>>>>                              >>>uncomplicated;
>>>>                              >>>the hurdle is parsing the dREL
>>>>            script
>>>>                              as a method of the object.
>>>>                              >>>Unless working with python, I'm
>>>>            not
>>>>                              sure that adopting python syntax
>>>>                              >>>for DDLm/CIF
>>>>                              >>>is of any great benefit;
>>>>            likewise for
>>>>                              dREL.
>>>>                              >>>
>>>>                              >>>That said, I have yet to
>>>>            actually do
>>>>                              anything with DDLm, let alone
>>>>                              >>>dREL, so I may be
>>>>                              >>>well off the mark. But even if
>>>>            this
>>>>                              is the case, I suspect there
>>>>                              >>>will be non-python programmers
>>>>            out
>>>>                              >>>there that have cause to work
>>>>            with
>>>>                              CIF and similarly will see no
>>>>                              >>>obvious benefit in
>>>>                              >>>CIF sharing python syntax
>>>>            (especially
>>>>                              if it only adopts it for one
>>>>                              >>>set of delimiters at the
>>>>                              >>>data-source level).
>>>>                              >>>
>>>>                              >>>Cheers
>>>>                              >>>
>>>>                              >>>Simon
>>>>                              >>>
>>>>                              >>>
>>>>                              >>>
>>>>                              >>>From: Herbert J. Bernstein
>>>> 
>>>>>>>>>> <<mailto:<mailto:yaya@bernstein-plus-sons.com>yaya@bernstein-plus-sons.c
>>>> 
>>>> 
>>>> 
>>>> 
>>>>  om><mailto:yaya@bernstein-plus-sons.com>yaya@bernstein-plus-sons.com>
>>>>                              >>>To: Group finalising DDLm and
>>>>                              associated dictionaries
>>>> 
>>>>>>>>>> <<mailto:<mailto:ddlm-group@iucr.org>ddlm-group@iucr.org><mailto:ddlm-gr
>>>> 
>>>> 
>>>>                              oup@iucr.org>ddlm-group@iucr.org>
>>>>                              >>>Sent: Saturday, 15 January,
>>>>            2011
>>>>                              21:16:59
>>>>                              >>>Subject: Re: [ddlm-group]
>>>>                              Relationship asmong CIF2, STAR,
>>>>            CIF1 and
>>>>                              Python. .
>>>>                              >>>
>>>>                              >>>At 12:43 PM +0000 1/15/11,
>>>>            Brian
>>>>                              McMahon wrote:
>>>>                              >>>>It might be worth remarking
>>>>            (again)
>>>>                              that dREL is being developed as a
>>>>                              >>>>canonical methods description
>>>>                              language, and not necessarily the
>>>>                              runtime
>>>>                              >  >>>methods evaluator of choice
>>>>            for
>>>>                              future applications. It may be
>>>>            that in
>>>>                              >>>>practice future methods are
>>>>                              initially developed and most
>>>>            frequently
>>>>                              >>>>executed directly in Python or
>>>>            some
>>>>                              other language. As I see it, the
>>>>                              >>>>goal of CIF and DDL evolution
>>>>            is not
>>>>                              to exclude such a possibility.
>>>>                              >>>
>>>>                              >>>If we are trying to be Python
>>>>                              friendly and much of dREL is
>>>>            derived
>>>>                              >>>from a Jython implementation, I
>>>>            don't
>>>>                              understand why we are not
>>>>                              >>>conforming dREL, DDLm and CIF2
>>>>            to
>>>>                              Python conventions as closely as
>>>>                              >>>possible.
>>>>                              >>>
>>>>                              >>>
>>>>                              >>>
>>>>                              >>>
>>>>                              >>>
>>>>                              >>>
>>>>                              >>>
>>>>                              >>>At 12:43 PM +0000 1/15/11,
>>>>            Brian
>>>>                              McMahon wrote:
>>>>                              >>>>On Thu, Jan 13, 2011 at
>>>>            05:35:21PM
>>>>                              -0600, Bollinger, John C wrote:
>>>>                              >>>>>
>>>>                              >>>>>  (snip)
>>>>                              >>>>>
>>>>                              >>>>>  CIF2 <=> CIF1:
>>>>                              >>>>>  To the greatest extent
>>>>            feasible,
>>>>                              well-formed CIF1 documents should
>>>>            be
>>>>                              >>>>>  well-formed CIF2 documents
>>>>                              (modulo a CIF version
>>>>            identification
>>>>                              >>>>>  signature) having the same
>>>>                              meaning.
>>>>                              >>>>
>>>>                              >>>>Agreed.
>>>>                              >>>>
>>>>                              >>>>>  CIF2 <=> STAR:
>>>>                              >>>>>  Inasmuch as CIF1 is derived
>>>>            from
>>>>                              STAR, I think it appropriate for
>>>>                              CIF2
>>>>                              >>>>>  to look first to STAR,
>>>>            including
>>>>                              its post-CIF1 development, for new
>>>>                              >>>>>  features it may need.  Even
>>>>            if
>>>>                              CIF2 is not 100% compatible with
>>>>            STAR,
>>>>                              it
>>>>                              >>>>>  is worthwhile to avoid
>>>>            diverging
>>>>                              without compelling reason.
>>>>                              >>>>
>>>>                              >>>>Agreed
>>>>                              >>>>
>>>>                              >>>>>  CIF2 <=> Python:
>>>>                              >>>>>  I see no particular reason
>>>>            for
>>>>                              any formal relationship here
>>>>            beyond
>>>>                              >>>>>  Python's role as the
>>>>            indirect
>>>>                              inspiration for CIF2's new
>>>>                              >>>>>  triple-quote syntax.  I am
>>>>            wary
>>>>                              of the idea of tying CIF tightly
>>>>            to
>>>>                              >>>>>  a particular language.
>>>>            CIF2
>>>>                              documents are not and never will
>>>>            be
>>>>                              >>>>>  Python programs.  I could
>>>>            imagine
>>>>                              embedding Python in CIF or vise
>>>>                              >>  >>>  versa, but I have seen no
>>>>                              evidence to suggest that greater
>>>>                              similarity
>>>>                              >>>>>  between the two languages'
>>>>            syntax
>>>>                              and semantics would benefit
>>>>            efforts
>>>>                              >>>>>  such as those.
>>>>                              >>>>
>>>>                              >>>>Agreed. As I mention
>>>>            elsewhere,
>>>>                              there is a greater influence on
>>>>            the
>>>>                              >>>>prototype dREL (arising from
>>>>            the
>>>>                              initial Jython implementation),
>>>>            and
>>>>                              >>>>the list and table data types
>>>>                              doubtless arise from that also.
>>>>                              >>>>
>>>>                              >>>>It might be worth remarking
>>>>            (again)
>>>>                              that dREL is being developed as a
>>>>                              >>>>canonical methods description
>>>>                              language, and not necessarily the
>>>>                              runtime
>>>>                              >>>>methods evaluator of choice
>>>>            for
>>>>                              future applications. It may be
>>>>            that in
>>>>                              >>>  >practice future methods are
>>>>                              initially developed and most
>>>>            frequently
>>>>                              >>>>executed directly in Python or
>>>>            some
>>>>                              other language. As I see it, the
>>>>                              >>>>goal of CIF and DDL evolution
>>>>            is not
>>>>                              to exclude such a possibility.
>>>>                              >>>>
>>>>                              >>>>Regards
>>>>                              >>>>Brian
>>>> 
>>>>            >>>>_______________________________________________
>>>>                              >>>>ddlm-group mailing list
>>>> 
>>>>>>>>>>>> <mailto:<mailto:<mailto:ddlm-group@iucr.org>ddlm-group@iucr.org><mailto
>>>> 
>>>> 
>>>> 
>>>> :ddlm-group@iucr.org>ddlm-group@iucr.org><mailto:<mailto:ddlm-group@iucr.or
>>>> 
>>>> 
>>>> 
>>>> 
>>>>  g>ddlm-group@iucr.org><mailto:ddlm-group@iucr.org>ddlm-group@iucr.org
>>>> 
>>>>>>>>>>>> <<<http://scripts.iucr.org/mailman/listinfo/ddlm-group>http://scripts.i
>>>> 
>>>> 
>>>> 
>>>> ucr.org/mailman/listinfo/ddlm-group><http://scripts.iucr.org/mailman/listin
>>>> 
>>>> 
>>>> 
>>>> fo/ddlm-group>http://scripts.iucr.org/mailman/listinfo/ddlm-group><<http://
>>>> 
>>>> 
>>>> 
>>>> scripts.iucr.org/mailman/listinfo/ddlm-group>http://scripts.iucr.org/mailma
>>>> 
>>>> 
>>>> 
>>>> n/listinfo/ddlm-group><http://scripts.iucr.org/mailman/listinfo/ddlm-group>
>>>> 
>>>> 
>>>> 
>>>>            http://scripts.iucr.org/mailman/listinfo/ddlm-group
>>>>                              >>>
>>>>                              >>>
>>>>                              >>>--
>>>> 
>>>>            >>>=====================================================
>>>>                              >>>  Herbert J. Bernstein,
>>>>            Professor of
>>>>                              Computer Science
>>>>                              >>>    Dowling College, Kramer
>>>>            Science
>>>>                              Center, KSC 121
>>>>                              >>>        Idle Hour Blvd,
>>>>            Oakdale, NY,
>>>>                              11769
>>>>                              >>>
>>>>                              >>>
>>>>            +1-631-244-3035
>>>>                              >>>
>>>> 
>>>>>>>>>> <mailto:<mailto:<mailto:yaya@dowling.edu>yaya@dowling.edu><mailto:yaya@d
>>>> 
>>>> 
>>>> 
>>>> owling.edu>yaya@dowling.edu><mailto:<mailto:yaya@dowling.edu>yaya@dowling.e
>>>> 
>>>> 
>>>> 
>>>>            du><mailto:yaya@dowling.edu>yaya@dowling.edu
>>>> 
>>>>            >>>=====================================================
>>>> 
>>>>            >>>_______________________________________________
>>>>                              >>>ddlm-group mailing list
>>>> 
>>>>>>>>>> <mailto:<mailto:<mailto:ddlm-group@iucr.org>ddlm-group@iucr.org><mailto:
>>>> 
>>>> 
>>>> 
>>>> ddlm-group@iucr.org>ddlm-group@iucr.org><mailto:<mailto:ddlm-group@iucr.org
>>>> 
>>>> 
>>>> 
>>>> 
>>>>  >ddlm-group@iucr.org><mailto:ddlm-group@iucr.org>ddlm-group@iucr.org
>>>> 
>>>>>>>>>> <<<http://scripts.iucr.org/mailman/listinfo/ddlm-group>http://scripts.iu
>>>> 
>>>> 
>>>> 
>>>> cr.org/mailman/listinfo/ddlm-group><http://scripts.iucr.org/mailman/listinf
>>>> 
>>>> 
>>>> 
>>>> o/ddlm-group>http://scripts.iucr.org/mailman/listinfo/ddlm-group><<http://s
>>>> 
>>>> 
>>>> 
>>>> cripts.iucr.org/mailman/listinfo/ddlm-group>http://scripts.iucr.org/mailman
>>>> 
>>>> 
>>>> 
>>>> /listinfo/ddlm-group><http://scripts.iucr.org/mailman/listinfo/ddlm-group>h
>>>> 
>>>> 
>>>> 
>>>>            ttp://scripts.iucr.org/mailman/listinfo/ddlm-group
>>>>                              >  >>
>>>>                              >>>
>>>> 
>>>>            >>>_______________________________________________
>>>>                              >>>ddlm-group mailing list
>>>> 
>>>>>>>>>> <mailto:<mailto:ddlm-group@iucr.org>ddlm-group@iucr.org><mailto:ddlm-gro
>>>> 
>>>> 
>>>>                              up@iucr.org>ddlm-group@iucr.org
>>>> 
>>>>>>>>>> <<http://scripts.iucr.org/mailman/listinfo/ddlm-group>http://scripts.iuc
>>>> 
>>>> 
>>>> 
>>>> r.org/mailman/listinfo/ddlm-group><http://scripts.iucr.org/mailman/listinfo
>>>> 
>>>> 
>>>> 
>>>>           
>>>>  /ddlm-group>http://scripts.iucr.org/mailman/listinfo/ddlm-group
>>>>                              >>
>>>>                              >>
>>>>                              >>--
>>>> 
>>>>            >>=====================================================
>>>>                              >>  Herbert J. Bernstein,
>>>>            Professor of
>>>>                              Computer Science
>>>>                              >>    Dowling College, Kramer
>>>>            Science
>>>>                              Center, KSC 121
>>>>                              >>        Idle Hour Blvd, Oakdale,
>>>>            NY,
>>>>                              11769
>>>>                              >>
>>>>                              >>
>>>>            +1-631-244-3035
>>>>                              >>
>>>> 
>>>>>>>> <mailto:<mailto:yaya@dowling.edu>yaya@dowling.edu><mailto:yaya@dowling.ed
>>>> 
>>>> 
>>>>                              u>yaya@dowling.edu
>>>> 
>>>>            >>=====================================================
>>>> 
>>>>            >>_______________________________________________
>>>>                              >>ddlm-group mailing list
>>>> 
>>>>>>>> <mailto:<mailto:ddlm-group@iucr.org>ddlm-group@iucr.org><mailto:ddlm-grou
>>>> 
>>>> 
>>>>                              p@iucr.org>ddlm-group@iucr.org
>>>> 
>>>>>>>> <<http://scripts.iucr.org/mailman/listinfo/ddlm-group>http://scripts.iucr
>>>> 
>>>> 
>>>> 
>>>> .org/mailman/listinfo/ddlm-group><http://scripts.iucr.org/mailman/listinfo/
>>>> 
>>>> 
>>>> 
>>>>            ddlm-group>http://scripts.iucr.org/mailman/listinfo/ddlm-group
>>>>                              >>
>>>>                              >>
>>>> 
>>>>            >>_______________________________________________
>>>>                              >>ddlm-group mailing list
>>>> 
>>>>            >><mailto:ddlm-group@iucr.org>ddlm-group@iucr.org
>>>> 
>>>>>>>> <http://scripts.iucr.org/mailman/listinfo/ddlm-group>http://scripts.iucr.
>>>> 
>>>> 
>>>>                              org/mailman/listinfo/ddlm-group
>>>>                              >
>>>>                              >
>>>>                              >--
>>>> 
>>>>            >=====================================================
>>>>                              >  Herbert J. Bernstein, Professor
>>>>            of
>>>>                              Computer Science
>>>>                              >    Dowling College, Kramer
>>>>            Science
>>>>                              Center, KSC 121
>>>>                              >        Idle Hour Blvd, Oakdale,
>>>>            NY,
>>>>                              11769
>>>>                              >
>>>>                              >                  +1-631-244-3035
>>>>                              >
>>>> 
>>>>            <mailto:yaya@dowling.edu>yaya@dowling.edu
>>>> 
>>>>            >=====================================================
>>>> 
>>>>            >_______________________________________________
>>>>                              >ddlm-group mailing list
>>>> 
>>>>            ><mailto:ddlm-group@iucr.org>ddlm-group@iucr.org
>>>> 
>>>>>> <http://scripts.iucr.org/mailman/listinfo/ddlm-group>http://scripts.iucr.o
>>>> 
>>>> 
>>>>                              rg/mailman/listinfo/ddlm-group
>>>>                              >
>>>>                              >
>>>> 
>>>>            >_______________________________________________
>>>>                              >ddlm-group mailing list
>>>>                              >ddlm-group@iucr.org
>>>> 
>>>>            >http://scripts.iucr.org/mailman/listinfo/ddlm-group
>>>> 
>>>> 
>>>>                              --
>>>> 
>>>>            =====================================================
>>>>                                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
>>>> 
>>>>            =====================================================
>>>> 
>>>>            _______________________________________________
>>>>                              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
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 
>>>>    ____________________________________________________________________
>>>> 
>>>> _______________________________________________
>>>> 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
>>> 
>>> 
>> 
>> 
>> 
>> -- 
>> 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]
International Union of Crystallography

Scientific Union Member of the International Science Council (admitted 1947). Member of CODATA, the ISC Committee on Data. Partner with UNESCO, the United Nations Educational, Scientific and Cultural Organization in the International Year of Crystallography 2014.

International Science Council Scientific Freedom Policy

The IUCr observes the basic policy of non-discrimination and affirms the right and freedom of scientists to associate in international scientific activity without regard to such factors as ethnic origin, religion, citizenship, language, political stance, gender, sex or age, in accordance with the Statutes of the International Council for Science.