Seth Godin is a mentor to many marketers out there today. I’ve read a number of his books over the years and he has many comments. He is a self-professed “idea-giver” …read his latest blog posting. I specifically like his comment “ideas are easy, doing stuff is hard”. How true that is. Over the years I’ve had lots of ideas. I’ve shared many “beverage-based conversations” where big ideas have been put out. The trick is in the “money where your mouth is” execution of these ideas. Over the years I’ve had the pleasure of working with people who tend to deliver as well as talk. WAY more motivating than just listening to the promises of what could be.
A few years ago at a meeting in Washington I sat in on probably the earliest public forum discussion on the potential of InChI. As a result of excellent teamwork between NIST and IUPAC, and doing rather than just talking they got it done. There was some negativity expressed during the initial meetings about InChI but it did not distract the team from producing the prototype versions, initial release and now the latest update with InChIKey support.
Now, I’ll guarantee that Seth Godin doesn’t know what an InChIKey is (Seth, if you’re reading this prove me wrong 🙂 ). But I want to take the position of supporting the Big Idea of structure searching the web and suggesting InChI key as one way execute on this now. There is a lot of passion around doing this and it has shown up in a number of postings by Rich, by Joerg (in regards to Wikipedia in this discussion), by Egon (discussing RDF’ing molecular space) and Jim, among others.
I am reading and hearing exchanges about the web being made structure searchable and my mind drifts immediately to the “it’s not enough” stance. The InChIKey should address some of the issues seen with InChI string searches and likely will be way more popular with the search engines. As commented last night on ChemSpider news the InChI keys on ChemSpider now link directly to a Google search.
The challenge remains, once all of those keys are out there how will the web be SUBstructure searchable or SIMILARITY searchable. The solution would appear to be a centralized repository of structures with their associated InChI strings and InChIKeys. The InChIKey cannot be reversed to the structure. A centralized repository of millions of structures and associated InChI strings and keys would allow that repository to be searched by substructure/similarity and then when a structure(s) of interest is identified then the Google search on that string/key could be kicked off. Maybe the discussion regarding the creation of such a centralized repository has happened already so I’d be interested in hearing what the path forward for that is. If it’s happening then the questions are who will host, how will it be funded, is there a timeline etc. If it’s not happening or is way in the future then I have an interest in opening the discussion regarding using the ChemSpider database and appropriate services (presently under development) to provide an interim service.
Structure searching of the web is of course going to provide high value. It should not stop there of course. let’s have the proactive dialog now about the next phase to facilitate substructure and similarity searching. If the conversations are going on elsewhere please post the links as comments so that the readers can follow them. I’m sure that Egon, Joerg, Rich, PMR will all have thoughts about how this should look. The bottom line out there is if this is the path the underlying system needs to be able to handle at least 25 million structures (ChemSpider has 17 million already) in the short term and be scalable to many tens of millions. There aren’t too many open platforms that can do that yet. I am aware of commercial platforms supporting many millions but no Open Source platforms yet…