[Cod-bugs] "crystal structure alarm" / feature suggestion

[Behrnd Norwid]

        Dear developers and maintainers of COD,
    the other day I set an alarm in Thomson-Reuter's / Clarivate's literature data base Web of Science, i.e. periodically, the data base will check if an article interesting for me was cited by an other publication, and inform me -- even if the result shall be ``none'' (when the data base is not aware about a recent additional citation of this work).
    From this experience, on one side, and remembering that at least a considerable fraction of the data deposited in the COD may be retrieved by pattern comparison of Smiles which in turn may be automated, the thought occured to me that something similar might be implemented in the COD's portfolio.  There are two kind of alerts I am aware of:
+ The COD user wants to stay abreast about the discovery of new polymorphs of organic molecules, with models already curated in the COD.  This may be of interest for pharmacological relevant ingredients since at least some of the exhibit different physicochemical / pharmacocinetical properties and are patented / deplyed as "polymorph A" rather than "polymorph B".
+ To the astonishment of a COD user, there is a gap -- a crystalline compound was not yet characterized by diffraction analysis, or at least, COD is not (yet) aware of such a data set.  Beside the option to perform this characterization by himself / herself, s/he might be interested to learn when this characterization was done by someone else; and subsequently access the original publication COD points to.

    While I do not assume, that this list of reasons for such an alert is exhaustive, the two perspectives are appealing enough to suggest such an implementation in crystallographic data bases, too.  Provided COD is updated much more frequently than CCDC's CSD file, perhaps the implementation shows of benefits earlier, than at CCDC to which I wrote, too.  From which specialty within chemistry COD receives the most data per unit of time?  This likely different ratio might contribute to determine if such crawling across the newly curated data shall be performed monthly, quarterly, once a year.  On the other hand, maybe the task is facilitated because COD is backed by subversion; and the repeatitive crawling "just" has to be told to go back in history to the date of the last crawl per profile (in other words, commit #212121), instead of digesting all folder and data again.
    With kind regards,
Norwid Behrnd, PhD

 
|  | Virenfrei. www.avg.com  |

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.crystallography.net/pipermail/cod-bugs/attachments/20180130/83f7776c/attachment.html>