It has already been done and posted.
http://syngen2.chem.brandeis.edu/syngen.html (http://syngen2.chem.brandeis.edu/syngen.html)
http://syngen2.chem.brandeis.edu/sample-output/lysergic_A106.html (http://syngen2.chem.brandeis.edu/sample-output/lysergic_A106.html)
I'm not fat just horizontally disproportionate.
> because some chems aren't readily available-
Okay, but isn't that just supposed to be a cue which inspires chemists to make their own?
> that and you need to have some knowledge of how a reaction will proceed before attempting it. in other words, you need an analogous reaction written up in a journal (or other verifiable reference) for each and every step.
But doesn't the program itself contain these references for each reaction it proposes?
> this wouldn't be nearly as important if the chemicals for such novel syntheses were readily available by the liter/kilogram at low prices.
The intermediate reagents shown on the syngen-page seem to be rather small - surely they could be synthesized inexpensively?
> however, they're not and therefore nobody wants to dick around with something that's not even likely to work.
I just thought most of these programs were written so they almost always choose reactions which are well-described, and minimized actions which could threaten to influence the predictability of the reaction.
This is what we (I) need: http://www.sciencebase.com/dec98_iss.html (http://www.sciencebase.com/dec98_iss.html)
A neural network to spot where there could be a conflict in a proposed reaction.