Comments on: Adding Publication Details to Chemical Structures on ChemSpider

By: will

will — Fri, 29 Feb 2008 13:32:41 +0000

Structures are facts and can’t be copyrighted, regardless of whether the organisation providing the feed thinks that they can be. This, of course, probably does not apply to:
- images of structures (i.e. creative works)
- the contents of the file containing the structure e.g. SDF, CIF etc. (i.e. the expression of data)

By: Antony Williams

Antony Williams — Thu, 28 Feb 2008 13:36:42 +0000

David…you are correct that a couple of minutes to every structure would be time-consuming…but that’s the nature of Wikipedia contributions of course.

Now, is there a way to automate the process. Sure, we could scrape abstracts and titles and even figure out how to link up structures to names in an automated process but the quality would be the issue. How to confirm that the structure-publication association is correct. I think that manual association will certainly offer higher quality. Now, if publishers were allowing InChIs to be embedded then this would be much easier. We could, in theory, deal with RSS feeds with InChIs included and connect those up. What do you think re. copyright? If something is being RSS fed with structures and article titles and partial abstracts is it acceptable to link up to ChemSpider? I think the answer should be yes but….

By: David Bradley

David Bradley — Wed, 27 Feb 2008 18:38:59 +0000

That’s a stupendous structure, with those three unconnected oxygens in a ring-arrangement.

Meanwhile, I guess you have to remember that 60 seconds here, 60 seconds there all mounts up. I suspect there’s no way to automate the process, I guess that’s the whole point that there isn’t…