Discussion List Archives

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

Minutes of 2021 COMCIFS meeting

  • Subject: Minutes of 2021 COMCIFS meeting
  • From: Brian McMahon via cif-developers <cif-developers@xxxxxxxx>
  • Date: Tue, 21 Sep 2021 10:39:04 +0100
  • Cc: Brian McMahon <bm@xxxxxxxx>
Committee for the Maintenance of the CIF Standard (COMCIFS)
Meeting Online 31 August 2021 12:00 UTC
Present: James Hester* (Chair), Herbert Bernstein*, John Bollinger*,Saulius Grazulis, Mike Hoyland, James Kaduk, Brian McMahon*, AndriusMerkys, Peter Murray-Rust, Brian Toby, Antanas Vaitkus, JohnWestbrook*, Simon Westrip
* Voting members
1. COMCIFS report last triennium--------------------------------
The triennial report for 2017-2021 (a four-year period on account ofthe postponement of the IUCr XXV Congress) was circulated beforehandand will be published as part of the Congress Report in ActaCrystallographica Section A.
2. Any questions/discussion of report-------------------------------------
There were no comments. Among the items noted were that submissions tothe wwPDB in CIF format were now mandatory, that work on the secondedition of International Tables Volume G was progressing, thatcollaboration between COMCIFS and the NeXus International AdvisoryCommittee (NIAC) continued to be fruitful, and that there was contactwith European efforts around materials modelling (EMMO and OPTIMADE).
3. Actions arising from CommDat meeting 25th August---------------------------------------------------
COMCIFS works closely with the IUCr Committee on Data (CommDat),and noted the following activities discussed at the CommDatpost-Congress meeting.
The high-pressure community have been working for some time on a CIFdictionary, and are keen to make progress with this.
The quantum crystallography community also sees a need to capturedetails of non-spherical atomic form factors in a CIF format (asevidenced by Dylan Jayatilaka's keynote talk at the Congress and apresentation by Simon Grabowsky and Krzysztof Wozniak at thepre-Congress CommDat workshop on chemical crystallography).
The Commission on Powder Diffraction held an online meeting after theCongress to discuss further developments of the pdCIF dictionary andapplications. Their priority is to facilitate validation of powderstructure determinations submitted to IUCr journals, but they also seea need to capture all details of serial refinements across ranges oftemperature and pressure; and the desirability of working withinstrument vendors to improve the capture of raw data direct from thediffractometer to CIF.
The Commission on Electron Crystallography would also like to see CIFdictionary extensions to capture special aspects of electrondiffraction. Many members of the Commission also work with nanoED, aEuropean academic consortium that has agreed placements at the Chesteroffice with two students or postdocs to learn about the formalaspects of CIF.
It was also noted that links to CIF dictionaries, templates andexample files on the IUCr web site used ftp-based URIs. These were nolonger fetched by current web browsers. Brian McMahon (BM) undertookto replace these by http(s)-based URLs, and would review the best wayto provide location information in the CIF dictionary register.
4. Revitalising COMCIFS-----------------------
4.1 Discussion of ideas-----------------------
While acknowledging that the mmCIF/PDBx family of dictionaries waskeeping track effectively with developments in biologicalmacromolecular structural science, James Hester (JH) reflected thatother areas under COMCIFS supervision had plateaued for much of thepast decade. He felt that there was a need for COMCIFS to be, and tobe seen to be, more productive in extending CIF standards to noveltechniques and scientific growth areas. In addition to the initiativesdiscussed by CommDat, he gave neutron diffraction and the squeezealgorithm as examples of areas where new definitions and procedures wererequired. He did note that dictionaries were now being developed onGitHub, permitting access to a larger group of people. His earlysuggestion (on the COMCIFS email list) that it might be useful to havea technical advisory committee separate from the dictionary contentdevelopers had not found favour - it was generally felt that technicalimplementation and domain knowledge should be intimately coupled. Hedid identify the problem that the machinery of COMCIFS was opaque tothe outside world, and that some mechanism was needed to make it aseasy as possible for people to engage with COMCIFS activities when newstandards were desired. He invited further input on ways to revitalisethe work of the Committee.
Herbert Bernstein (HJB) argued that the most important step was toprovide clear step-by-step instructions (on the IUCr web or in print)on how to go about creating a new dictionary. He also suggestedconsidering the ANSI/ISO model where standards creators are required toreview the status of the standard on a fixed cycle; but emphasisedthat the engagement of stakeholders was more important than thespecific process adopted.
Peter Murray-Rust (PMR) considered the CIF dictionary effortoutstanding, and used it as a model in his current molecular and plantscience activities. He felt it was important to identify "hot spots"where people were actively developing ontologies, and saw materialsscience as an area where there was much current activity. He alsorecommended getting CIF identifiers into WikiData, and has alreadyworked with Saulius Grazulis (SG) and Antanas Vaitkus (AV) on puttingCOD identifiers there. This would put crystallography in front of manymore people through the distributed nature of WikiData. JH mentionedthe European Materials Modelling Council (EMMC) as a materialsscience initiative that he had engaged with, and PMR referred tothe BIG-MAP (Battery Interface Genome – Materials Acceleration Platform)and Material Genome projects, but pointed out that such initiativestended to flourish during a period of supported funding, but hadlimited longevity.
SG reported that he had had very positive experiences with GitHub andsimilar collaborative platforms in developing standards (OPTIMADE) andin software development, and that the use of a familiar communityplatform could allow IUCr to attract other groups with an interest indeveloping standards to work within a common development environment.[JH demonstrated some aspects of the existing COMCIFS GitHubimplementation, which allowed this sort of discussion and peerreview for the current suite of CIF dictionaries and their conversionto DDLm.] If COMCIFS were seen to be inclusive towards new ideas,that would encourage other groups to look to COMCIFS for authority,consistency and expert guidance. JH agreed on the merits of GitHub,but saw a need for documentation to help new users to use it to bestadvantage. He also made the point that as GitHub used version control,it was possible to allow experimentation by anyone interested incontributing, thus creating a more inclusive and welcoming environment.
JH reviewed some ideas he had presented to the IUCr Executive Committee:
* Topic-focused virtual workshops* IUCr contribution to conference/sabbatical for completion of CIF work* Lightweight newsletter for mailing lists (e.g. quarterly, with details   of recent GitHub activity)* Training modules for dictionary authors* Formalised and published governance procedures* Commissions take responsibility for dictionaries
He now thought that simple guidelines and how-to's might usefully takethe place of unduly formalised governance procedures.
HJB cautioned against a system that relied solely on Commissions,because of the danger of activities behind closed doors that couldstifle wider community input. JH argued that the Commission was atleast answerable to the Executive Committee. But there was generalagreement that any development activity should be carried out on aplatform such as GitHub, where openness was a feature and anyinterested contributors from the wider community could provideinput. It was felt that COMCIFS would liaise with a sponsoringCommission and would provide guidance on setting up a GitHub sitelinked within the framework of COMCIFS activities.
BM pointed out that there was a historic diversity of dictionary-buildingprojects - some Commissions (Aperiodic Crystals, Magnetic Structures)had proven very effective; other dictionaries (electron density,topology) had arisen from focussed groups of individualprogrammers. This suggested some ongoing flexibility in thecomposition of dictionary working groups; but Commission liaison wasworthwhile where that seemed useful or appropriate; and the idea of acommon family of GitHub sites would give a more complete andcoherent picture of overall activities. He also suggested (1) thatjournal editors should also be brought into working groups to design anew table of mandatory experimental data items that would inform thepublication requirements to be met by a new dictionary; and (2) someform of roadmap be published on the CIF website to show what areas ofstructural science had existing CIF dictionaries, where these wereunder revision or construction, and what areas were barren and hadpotential for future ontology development.
John Westbrook (JW) agreed that publishers (and relevant repositories)were key players who should be involved in new developments. Anotheraspect of diversity was the variety of stakeholders who might havedifferent requirements as a new area of CIF was developed, and so itwas essential that any working groups sought to engage with andrepresent all of these diverse (and sometimes conflicting)requirements. He made the point that (from a repository viewpoint)collecting the data was dependent on the software actually in usewithin the community that was going to support the desired standards.
JW also emphasised that the experience of the mmCIF community wasthat greatest productivity flowed from small or medium-sized workinggroups able to work together on a regular basis. Asynchronous(email-based) discussion was not very productive where complexrequirements needed to be understood and agreed. A model where peoplecan be brought together, focus on the immediate requirements, andrevisit progress on a timescale where enough time has been allowed todevelop code, but not so much that details get forgotten, has workedeffectively.
SG thought the idea of a lightweight newsletter for developers a goodone, but pointed out the potential usefulness of the the IUCr Newsletterfor periodically carrying news on data standardisation to a much widercommunity. He also pointed out that email-led development could workwell, provided the discussion was led and directed by a skilled projectleader. Returning to the problems faced by groups new to CIF ofgetting up to speed on how to create a dictionary, he recommended thatCOMCIFS publish contact details of members who could act as mentors topeople coming freshly to the field.
Jim Kaduk (JK) reviewed the status of the powder dictionary. While theimmediate priority was to encourage more authors to submit structuredeterminations based on powder to IUCr journals (through fine-tuningof checkCIF procedures), he felt the pdCIF dictionary needed to bereviewed so as better to handle CIFs with mixtures and/or quantummechanical calculations. It was also the case that parametricexperiments were much more important than when the pdCIF dictionarywas first published, and new experimental metadata were needed todescribe current practice. He volunteered to explore these issues. JHsuggested this could be the basis of a focussed workshop. JK indicatedthat authors of the most widespread packages (GSAS, TOPAS, FullProf)would be among the stakeholders that one needed to engage.
4.2 Next steps--------------
JH would consider the discussion and circulate to the group asummary of suggestions on moving forward.
BM would update ftp: based links on the IUCr website.
BM would think about a suitable visual indication of establishedCIF ontologies, ones under development, and areas requiring newstandards.
5. Any other business---------------------
None was raised.
The meeting concluded at 13:00 UTC.

Brian McMahonSecretary_______________________________________________cif-developers mailing listcif-developers@iucr.orghttp://mailman.iucr.org/cgi-bin/mailman/listinfo/cif-developers

Reply to: [list | sender only]