Metabolomics: ChemSpider - API & "Mass Spectrometry and Metabolomics Meeting in Research Triangle Park"

This regards an older post from the Chemspider blog:

The metabolomics community presently utilizes public domain databases such as KEGG, LipidMap, DrugBank and a myriad of other online resources to assist in the analysis of data. However, rich as these resources are, they are limited in scope, are challenged by known data quality issues, and are not directly focused on serving the needs of the metabolomics community. ChemSpider is an online resource for the chemistry community hosted by the Royal Society of Chemistry with the intention of linking together online chemistry resources, cleaning and curating chemistry related data and collectively serving a number of communities. ChemSpider has been used by members of the mass spectrometry community, including instrument vendors, for the past 3 years. This roundtable meeting is to provide an overview of how ChemSpider is presently used by scientists working in the domain of metabolomics and garner feedback from the existing user base as well as new potential users to help define how ChemSpider can be enhanced to further support the needs of this community.

ChemSpider, Mass Spectrometry and Metabolomics Meeting in Research Triangle Park

Whilst Chemspider is without doubt, amongs the better chemistry-related Web 2.0 data portals out there, it is unfortunately not an open source effort. Looking at its own web-site code reveals a patchwork of several web-developers. Less disparity between Chemspider's programmers would lead to a more streamlined code, for faster page refreshes. Their implementation is based on jquery and YUI amongst others.

The issues I faced with Chemspider is threefold:
Firstly there is their API, which seems to be a bit of a let down on the speed side, for reasons unknown to me. The APIs are not well documented, and partially don't seem do work or at least Chemspider's documentation has been rendered obsolete due to internal updates.
examples are provided like this (in the mass spec API)

Curiously enough the examples aren't filled with actual....exemplary [sic]values, and no clue is provided as to the nature of these parameters, other than their suggested type.

For some reason, most of the provided form examples don't work. And I am still left with figuring out what the parameter rid (string) is. Unfortunately it isn't the chemical's IUPAC name.

Secondly, Chemspider does very little in allowing easy integration into third party applications. Amongst the examples provided is an over 6 year old C++ Windows MFC example. Mine, that old often don't work anymore because of obsolete Windows MFC APIs. In those 6 years, Microsoft has since then pushed its own .NET platform further away from MFC.NET platform further and away from MFC. In those 6 years Chemspider has also reinvented itself at least once.

Thirdly, and this for me is the biggest crux, Chemspider is not open source. The need is obviously there as development doesn't proceeds as fast as it could, nor is it easily adoptable to ongoing data aggregation and data crowdsourcing efforts.

At last, crowdsourced based data curation, is where Chemspider definitely triumphs!