Comments on: RSC and Southampton drive the chemical semantic web

By: Richard

Richard — Fri, 27 May 2011 11:57:27 +0000

It’s not a copyright issue. We measure the value of providing ChemSpider by it being used – while we’ve always been prepared to discuss this with individual collaborators, there are risks to us of providing a public data download as well as benefits.

And I still have to be convinced that the semantic web should depend on static downloads of fluid database resources being shipped around. Am happy to discuss further but might be better done offline.

By: Egon Willighagen

Egon Willighagen — Fri, 20 May 2011 08:48:32 +0000

“In a connected world where data changes, I don’t get the stipulation to provide a download in full.”

Can you elaborate? The need for such an index is clear to me, but I can get a group of people together showing interest in this index. Also, I am not sure what you refer to as a ‘download in full’. A full download of ChemSpider is not requested.

“And it’s making it difficult to apply more clarity about the licensing.”

Do you mean that people that provided ChemSpider with data claim copyright on the chemical structures (InChIs) they provided? Or, how is the licensing unclear? Isn’t ChemSpider/RSC sole owner of their CSID identifiers?

“If you want the right answer, now, why is the web service not the best answer rather than a download that’s out of date as soon as it’s generated?”

Because it does not make sense to query a webinterface iteratively for a set of compounds in some database. That would needlessly stress the ChemSpider web services.

By: Richard

Richard — Fri, 20 May 2011 07:46:45 +0000

In a connected world where data changes, I don’t get the stipulation to provide a download in full. And it’s making it difficult to apply more clarity about the licensing.

If you want the right answer, now, why is the web service not the best answer rather than a download that’s out of date as soon as it’s generated?

InChIToCSID

InChIKeyToCSID

By: Egon Willighagen

Egon Willighagen — Thu, 19 May 2011 16:28:58 +0000

The thing with a semantic web is that there needs to be a web. In and out. Out is easier, though ChemSpider is not linking out to other resources yet. Just using the InChI to http://rdf.openmolecules.net/ would be a start. In requires ChemSpider to make a table with two columns available for download (under CC0): column 1 containing the InChI, column 2 the ChemSpider ID. That way, other databases can point to the ChemSpider RDF. Until either or both are done, ChemSpider cannot be seen as part of the semantic web. And both are easy to do.

By: Richard

Richard — Thu, 19 May 2011 15:48:12 +0000

Missed out examples earlier…

Using the RDF Permalink via ChemSpiderID (CSID)

For example:

http://www.chemspider.com/Chemical-Structure.7787.rdf
Or:
http://www.chemspider.com/rdf.ashx?q=7787

Using a Search Term to access the Atom feed

For example:

http://www.chemspider.com/rdf.ashx?q=cyclohexane
Or:
http://rdf.chemspider.com/cyclohexane

*Note: When searching using a term other than ChemSpiderID (CSID) then an ATOM feed will be returned as multiple results must be handled. The ATOM feed contains links to the individual RDF Permalinks as well as a human-readable XHTML summary fragment.

By: Richard

Richard — Mon, 16 May 2011 20:21:56 +0000

Hello Egon! These are pretty much rhetorical questions, right? We’re trying to build practical solutions a step at a time – now we have something to build on. We wouldn’t have made it available if we didn’t want it to be freely used, and as a basis for further expansion and collaboration.

By: Egon Willighagen

Egon Willighagen — Mon, 16 May 2011 15:39:21 +0000

Hi all, congrats on this really interesting collaboration! I am intrigued by the ‘drive the chemical semantic web’ claim, and am most interested in learning how the RSC and Southamption ‘drive’ this community. Can you elaborate on how this effort increases the momentum of chemistry in RDF? Particular problems the community faces is non-Open data, unclear licensing, lack of downloads, etc. As such, I am wondering how you integrated 400 resources, and how you have overcome these problems. I am also most interested in learning how this collaboration will collaborate with the rest of the community, such as Bio2RDF, Chem2Bio2RDF, LODD, and others. I know press releases are not the place for details, nuances, etc, but perhaps add those as a reply to this comment? (Did I say yet I hate press releases in scientific context?)