Crystallography World Wide

WWW Publication of Structural Results


Posting from Phil Fanwick to sci.techniques.xtallography

From Phil Fanwick <fanwick@xray.chem.purdue.edu>
Newsgroups: sci.techniques.xtallography
Subject: Posting Structural Results on the WWW
Date: Thu Sep 21 16:03:58 MET DST 1995
Organization: Department of Chemistry, Purdue University
It has become popular for crystallographic groups to create home pages on the WWW. Frequently recent structural results are posted on these pages. A recent change in the policies at the American Chemical Society could result in such information being rejected by ACS journals.

On April 3, 1995 the JACS Editorial Advisory Board issued the following change The posting of work on an electronic bulletin board constitutes prior publication and such work cannot be accepted for publication in JACS. I recently requested clarification from a Senior Research Associate at the ACS and received the folowing response. The current Notice to Authors for JACS states 'the work reported therein has not received prior publication and is not under consideration for publication elsewhere in any medium, including electronic journals and computer data bases of a public nature.' The revision that you mentioned they plan to add of 'electronic bulletin boards' only makes it more clear that JACS won't print items previously posted.

I looked at the Notice to Authors for Inorganic Chemistry as these Notices vary somewhat from journal to journal. The IC notice just says that the contribution must be original and must not have appeared elsewhere. This is a little more vague than the JACS notice. However, when I brought up the idea that authors might post a 'structure of the week' on the Web with an editor, his response was I think an author would have to be crazy to put it out on the web prior to publication. So it looks like it is 'author beware' when posting this material.


Followup by Bev Vincent

From brv@msc.com (bev vincent)
Newsgroups: sci.techniques.xtallography
Subject: Re: Posting Structural Results on the WWW
Date: Thu Sep 21 21:00:17 MET DST 1995
Organization: Molecular Strucuture Corporation
This is intriguing. Actually, the editors are very correct to say that it would be ill-advised to post results VERY publicly prior to publication. However, what exactly constitutes 'publishing results'? If you just put up a picture of a crystal structure, there is probably not enough information contained in it to be considered crystallographic results.

If someone posted a picture of the structure of an interesting new compound, even with cell parameters, that would not provide me with enough info to either derive any information from the results (e.g. bond lengths and angles) or to help me solve the structure if I had diffraction data from the same sample. At best, I might be able to glean the atomic connectivity pattern, some gross conformational information, and the chirality. Only if the actual atomic coordinates were made available would it, IMHO, constitute some form of publication. A picture of the molecular structure in the form of a GIF or a PostScript image could equally well have been derived from a molecular mechanics calculation or from an artists rendering.

It will be interesting, though, to see if the journals feel the same way!


Followup by John C. Huffman

From "John C. Huffman" <:huffman@indiana.edu>
Newsgroups: sci.techniques.xtallography
Subject: Re: Posting Structural Results on the WWW
Date: Fri Sep 22 03:12:57 MET DST 1995
Organization: Indiana University Molecular Structure Center
The posting of results on a web site (or other electronic media) brings up several fascinating points. If I place data (detailed or not) on a server, how does one refer to it? Since there is no way to insure that data will remain, or that it will not change, it will be somewhat difficult to properly cite. Not only would you have to cite the electronic address, but also the date. If the data are subsequently removed or updated, how do we use it? There is no way to force a person to leave data, or to not change data which he is placing in an electronic form. One can argue that CCDC could (and perhaps should) scan for such data and place it in their database, but according to my legal people my copywrite prevents them from doing so until I give them permission.

This brings up several other interesting points. If you fear that someone else may be about to "scoop" you, simply place your data (complete with mistakes) on the bulletin board and claim prior authorship. Most journals insist they will not publish a structural result that has previously been published, even if the new results are markedly better. After leisurely finishing up the structure you can delete it from the bulletin board and submit it. If the editors claim that removing the data does not count (i.e. it is still published) they leave themselves open for all kinds of problems. All I have to do is put a note in each review that "this data was available electronically last year". How can they prove that it wasn't? In addition, how are the editors going to insure that a structure has not been placed on a web site or bulletin board? At last count there were already several million servers, and certainly there are thousands of bulletin boards that are publically available. If the editors claim that the ruling pertains only to the authors own data on his/her own web site, the absurdity of the policy will be obvious (i.e. have you given any "public" seminars in your own department where this data was disclosed?)

The unfortunate side to this is that scientists who would be willing to share data will become reluctant to do so. I would estimate that my laboratory probably has over 1000 "quality" (and admittedly some not so good) structures that have not been published, most of them low temperature data sets. In some cases the structure turns out to be another phase of a know compound, or simply a redetermination (accidental or on purpose) of a know phase at low temperature. Many of the organic compounds we do may not be published since they are merely confirmations of intermediates. This type of data might be of value to other researchers, but if I think that I am jepordizing a colleagues work, I certainly will be hesitant to make it available to others, with or without their permisssion. I can easily place passwork protection on all the data, but I personally feel that data which, for the most part, was obtained using public funding should be available to anyone who can make legitimate scientific use of it.

Computers are rapidly changing the way we do our research and, like it or not, the way we share our research results. I would urge the policy makers to think carefully before making protectionist rules that hurt the entire community.


Followup by Martin Kroeker

From martin@tutor.oc.chemie.th-darmstadt.de (Martin Kroeker)
Newsgroups: sci.techniques.xtallography
Subject: Re: Posting Structural Results on the WWW
Date: Fri Sep 22 12:16:27 MET DST 1995
Organization: Technische Hochschule Darmstadt
It might be interesting to compare this to the attitude of the same editors towards poster presentations. If one presents a poster, especially at some smaller local conference, the data will be seen by only a select number of researchers who happen to attend this particular meeting. Everybody else will have a hard time searching for the "Proceedings" or "Book of Abstracts", which may have consisted of a few stapled sheets and may or may not have contained information on the posters. I feel that this is the closest 'conventional' analogue to a Web page that not everybody knows about and that may change at the whim of its administrator.


Followup by Paul Doyle

From boyle@laue.chem.ncsu.edu (Paul D. Boyle)
Newsgroups: sci.techniques.xtallography
Subject: Re: Posting Structural Results on the WWW
Date: Fri Sep 22 16:12:51 MET DST 1995
Organization: North Carolina State University
I think John brings up many good points here. But of the questions that need to answered are what exactly constitutes "publishing", and how can the process of peer review be maintained in an electronically "loose" environment like the Web? On my Web site I have set up a system where clients can download a .zip file of thier structure from the Web page. My main motivation for setting up this system was to minimize typographical errors in the manuscript preparation process (something I think journal editors would encourage).

I do not consider this "publishing" because access to the subdirectories holding the results are password protected. Results which are not "publically" available are not "published". I also agree with Bev Vincent's opinion that placing a .gif file of a molecule on a Web page shouldn't constitute "publication". The picture is almost meaningless, particularly in terms of the quantitative information which is normally associated with crystal structure determinations (bond lengths, angles, etc.).

In my view, publishing something means (ideally) that enough information is made publically available so that someone else can repeat the experiment or calculation. In crystallography, the fundemental quantities needed are the unit cell parameters and the diffracted intensities. If some was given that information, that person could duplicate the "published" results - normally atomic parameters, and derived molecular geometric parameters. With the advent of cheaper disk space and fast computers and networks, it may be time to reconsider the importance of lists of Fobs or even raw data in the publication process. As you know, it was journals which discouraged the publication of Fobs/Fcalc's and ADP's to "save printing costs". That constraint is being bypassed in this new age of electronic publishing. This gives us the opportunity to more thoroughly review crystallographic data.

I think the stickier issue here is the peer review process. "Publishing" something on the Web, circumvents the whole peer review process. This is a bad thing. There are already too many wrong structures that make it into print these days. Encouraging publishing results on the Web would only exacerbate this problem. It is the nature of the "internet" to be decentralized. In general, I believe this to be a good thing. However, the scientific community needs peer review, and this requirement does not mesh well with the autonomous nature of the net.

This post got a lot longer than I thought it would, but to conclude: Putting non-quantitative information on the Web such as .gif's and the like should be allowed and even encouraged. However, it is not a good idea for someone to bypass peer review by publishing on the Web and then using that "publication" as some sort of claim, or worse have someone else reference those results gathered from the Web.


Followup by Joseph H. Reibenspies

  
From reibenspies@chmbp2.tamu.edu (Joseph H. Reibenspies)
Newsgroups: sci.techniques.xtallography
Subject: regarding structure presentation on the WEB
Date: Tue Sep 26 16:01:55 MET 1995
Organization: Chemistry Dept. Texas A&M Univ.
I have been following this subject for some time now. Several editors have stated that poster presentation on the web would be acceptable, if that work results in a full article (no communications). I would suggest that all crystallographic work should be published first before posting to the WEB.

With that said, what about post-publication work? Do you need to seek permission from the publisher to post a structural diagram? The structural coordinates, and thus any graphics presentation of those coordinates, are the property of the journal, since most journals insist that you turn over the copyright to them. Is posting your own published structures on the WEB an infringement of that copyright?

There seems to be many questions to be resolved.


Followup by Tony Linden

Path: rzunews.unizh.ch!NewsWatcher!user
From: alinden@oci.unizh.ch (Tony Linden)
Newsgroups: sci.techniques.xtallography
Subject: Re: Posting Structural Results on the WWW
Followup-To: sci.techniques.xtallography
Date: 22 Sep 1995 07:48:04 GMT
Organization: University of Zurich
The use of WWW or other electronic means is an attractive option for those who wish to make their data available to the world, either rapidly or because there is no desire to publish elsewhere. Others have alrady offered interesting comments. There are some important points to be considered:

The structural data posted on the web or a bulletin board has not been checked for errors, except (hopefully) by the person who has done the determination. There is no guarantee of the reliability of the information or that the work has even been done properly and carefully.

Data published through journals has been refereed and therefore one would hope that the work has been appropriately scrutinized and that blatant errors have been detected and corrected. Unfortunately the referees of many papers which contain a crystal structure as a very minor part of the whole often do not pay too much attention to the structural data since their expertise and interest probably lies in another area.

Data that appears in a database (e.g. CSD) has been checked for consistency by the database people and is therefore reliable - even if the quality of the data varies tremendously.

Publication on a web site or bulletin board can complicate the job of finding data. One might never being sure that the data isn't already out there SOMEWHERE! We have to be careful to at least centralise the information in some way.

It is not necessary to publish a structure in a journal for it to be entered in the CSD. They accept data as a "personal communication", so all those never-to-be-published, high quality structures of "uninteresting" products or intermediates that are lying around in many laboratories could (and should) be gathered together and sent off to the CSD, or ICSD. All that needs to be done beforehand is to get the permission of the scientist who prepared each compound (or the professor of the group) to allow the data to be made public in this way. The CSD now have a nice electonic form, which can be used to submit data by e-mail, either combined with a CIF, or not (for those old pre-CIF structures). I am sure that labs which have a good archiving system would have no difficulty submitting old data.

What is the purpose for putting data on web servers and what constitutes "publication" of the data? Full data (atomic coords, etc.) certainly does. Just a picture? This serves only to show that the structure has been done. In this case, what is the creator of the web site trying to do? Just show what nice structures they are turning out, sort of a listing of the lab's output (maybe initially intended primarily for use within a department or university), or to lay a firm claim to a "hot" structure?

What is the ACS ruling on results "published" at a conference either as a poster or a normal presentation? If it is OK to publish such material later on in an ACS journal, we are entering murky waters. I believe the bio sciences have a continuous poster session somewhere on the Web. The definition of "publication" might need to be sharpened up a bit to cater for such things.

Speed of publication is sometimes of great concern for "hot" work. Perhaps the threat of electonic publication superceding journal publications will get some of the slower journals to speed up their act. If one well-known international journal can do it in 3 months without camera-ready manuscripts or foregoing proper refereeing procedures, why do others require 9 months or longer?

Journals are the traditional means for scientists to communicate their work to others. But their main purpose is just that - a means of communication. For many years, journals (and conferences) have been the chief public forum for the dissemination of results. If technology now allows other means of communication, should we not embrace it? Fast communication can only benefit us all. Of course guidelines need to be developed to ensure the reliability and appropriateness of the information, just as refereeing does for journals.


Followup by Jim Silverton

From jvs@helix (James V. Silverton)
Newsgroups: sci.techniques.xtallography
Subject: Re: Posting Structural Results on the WWW
Organization: National Institutes of Health, Bethesda, MD
Date: Wed Sep 27 20:51:40 MET 1995
I think the previous postings and Tony Linden's views should be discussed widely. I think the essence of "publication" is that the data and conclusions are subject to review before presentation and are reasonably readily available. The abstracts of an ACA meeting do not meet these definitions and the ACS has accepted papers on subjects presented at meetings. If the papers at an ACA meeting were "reviewed" they were only done very lightly, mainly for format and suitability and the data was not readily available to all.

Another type of previous publication where the abstract of the paper becomes widely available is the conference with a published proceedings such as the Proceedings of the American Peptide Symposia. These are reviewed but I suspect again very lightly and are abstracted by Chem. Abstracts but papers on the topics were accepted for later publication in ACS journals. Possibly, the ACS has now changed its rules or else they were misinterpreted previously.

It is a very difficult decision to make but I consider that the Internet should only be used for staking a claim or asserting priority if sufficient supporting data is also presented and readily available. I think an author who presents work on the WWW solely as partial description like a diagram cannot claim priority and may have no recourse if it is used by someone else.

A possible solution is to have some central rapid reviewing system whose approval could be certified at the time of electronic publication. It should not be too difficult to institute but would certainly have to be agreed to by sufficient major scientific societies and would take careful organization to prevent delays. This last may be a real alternative method of publication. The review process need not take too long; many of us can send reviews back to an editor in less than two weeks as requested by, for example, J. Org. Chem. and J. Chem. Inf. Comp. Sci.


Followup by John Huffman to Tony Linden's comments

From "John C. Huffman" <huffman@indiana.edu>
Newsgroups: sci.techniques.xtallography
Subject: Re: Posting Structural Results on the WWW
Date: Fri Sep 29 02:30:25 MET 1995
Organization: Indiana University Molecular Structure Center
[Text in italics has been added by the Crystallography W3VL Editor to avoid lengthy repetition of Linden's text.]

With respect to data checking by journals and databases, the data have been checked for consistency, and therefore can be considered consistent. The CSD does, in fact, discover many errors, but almost all of these errors are due to sloppy crystallography and typographical errors. As I look at the corrections that have been made, most are incorrectly entered numbers. These errors are presumably decreasing as we improve the networking and data transfer techniques. The CSD, and indeed editors and/or reviewers can only assume the quality of the data that was collected. In fact, Richard Marsh is one shining example of truly examining data--he finds many errors which have readily passed editors, reviewers, and the CSD staff because they simply do not have the time or resources to perform the analysis Prof. Marsh provides.

Finding data on a bulletin board or web site is a real problem, but believe me, lots of people are working on it.

The CSD and ICSD will accept data for unpublished structures. I assume that this constitutes "publication" and would certainly preclude publication later.

What is that purpose of putting data on the web and what constitutes publication of the data is perhaps the most import point. By putting the data on the Web, it is available to an unbelievable number of reseachers, teachers, students, and "John Q. Public" at essentially no cost. Only a small fraction of the population has access to the CSD, the ICSD, NIST, etc. The number of individuals accessing the Web is exploding daily and will continue. For this reason alone I am for having as much data as possible available to those who quite frankly have paid for it--the taxpayers. One nice thing about our field is that the final result (i.e. a molecular structure properly presented) is not only science, but art as well! In these days of shrinking financial support anything we can do to interest the public in science is probably worthwhile.

As far as practical reasons for placing data from my laboratory on the net, there are many. Files can become available to the student or faculty member as soon as the structure has been solved. While most structures go from the "solved" to a final report in the hands of the student in only a few days, there are many problem structures that may require weeks. By having the preliminary data available, my colleagues can be using the information (even preparing the publication) while I finish the model and refinement without introducing errors due to haste. Even with the "simple" structures, a simple note to the student when the data are posted means that I do not have to rush the departmental staff to get reports duplicated and sent out. A more practical aspect is the eventual savings of a lot of trees--my lab used 300 reams of paper last year for reports that for the most part did not need to be in hard copy form. As stated earlier, I can easily put password protection on the data files, or make the files available only on the subnets within my department, but who loses in such a situation? I think we all do.

Modernising publication techniques is the crux of the matter. If new technologies allow new ways to communicate, we should not look for reasons to fault the technologies, but try to iron out difficulties that will certainly be encountered.

As I see it, journals still are, and will remain for some time, the way to properly document and present research results in a formal manner. The web is a convenient way to allow rapid "sharing" of results which may be of use to colleagues and others, as well as data which may not be published but which the researcher does not mind sharing. The data so obained should be considered on a par with the much discouraged "private communication" references since there is no way to insure its longevity or accuracy.


29th Sept. 1995 - Crystallography World Wide Editor - Acknowledgements