Re: [Jmol-users] Jmol 14.0.6 released

[Otis R. <osr...@ch...>]

Bob and Pino

> 
> Possibly. At least for known compounds. Does this fail gracefully? 
> 

NIST fails gracefully because they return a data not found graphic image. I've never encountered a NIST server down issue, so I'm not sure what happens there.

The biggest problem with SMILES to InChI is server down with response dependent on serve request timeout - i.e. dead air. That's where the page locking gets to be a problem.

I'm trying to set things up so that an InChI server down at Resolver automatically switches over to ChemSpider. I have that under control client side, but not server side. On the client side, I do a very quick AJAX touch on Resolver after page loading. If Resolver does not respond in 5 seconds, I switch to ChemSpider. I played around with some time intervals, and 5 seconds seems OK. I've had some good opportunities to test this recently!

> x = script("show chemical cas")  // but produces a list


I get a list also in some situations. The trick is to sort them and use the lowest numerical value. For simple compounds, the NIST hit rate is pretty good.

All in all, the approach works fairly well even though it is a limited database. Of course, our interest is fairly simple molecules.

Otis

--
Otis Rothenberger
ot...@ch...
http://chemagic.com