Hello Paul,
I've really enjoy reading your blog. You're onto
something by reviewing synthetic papers the way you
do.
Like you, I agree that an InChI on a Webpage is
unsightly and unnecessary. Unfortunately, despite many
attempts by myself and others, I know of no way to
unobtrusively get an intact, long InChI into a Webpage
without breaking the layout.
Awhile back, I followed the path you're taking now,
but with no success. My intent was to have google
index the InChI so that it could be found by the
InChIMatic service:
http://inchimatic.com
I've written about my little experiment here:
http://depth-first.com/articles/2007/03 ... -chemistry
Six months later, Google still hasn't indexed the
InChI I tried to sneak through by hiding it in my
article. Nor have any of the other major search
engines. The only thing that comes up are links to the
above article, which puts the InChI in plain sight (I
used the overflow property to prevent my page layout
from breaking -- still, even that has it's
problems...)
The problem is that Google fights a daily battle
royale against spammers and Search Engine Optimization
(SEO) tricksters. All of the things we can think of to
hide InChIs are also the same techniques used by these
lugnuts trying to sell junk. Some
would even say that trying to hide "meaningless"
content such as InChIs in your webpage will result in
a penalty in Google's results. I can't vouch for that
theory, but who knows? At the very least the content
is ignored during Google's indexing process.
In Firefox 2/Linux, I couldn't see the alt text by
mousing over the second image in your article. Nor in
Linux/Konqueror.
Recently, the InChI team has released software that
builds fixed-length InChIs using a technique called
hashing:
http://www.iupac.org/inchi/release102.html
It's a promising approach, but like InChI will depend
on widespread adoption to become useful.
Given the current problems, I stick by my conclusion
that the Web still isn't ready for chemistry. That's
not to say the solution isn't waiting to be discovered
- just that it hasn't been found and made public yet.
cheers,
Rich