I haven't played this in a while. I didn't even know they stopped updating researchnet and the comparison graphs. That's kind of disappointing. I wish they could find some sort of alternative hosting for that. Seems like an important bit of the game to just cut out entirely.
I'm glad to read some of his conclusions in that post mortem article. The game is great, but he could get a much wider audience for future titles if he finds a way to put the puzzles into a more inviting environment. You're already building a game around what is essentially programming, so making your aesthetic trappings look like chemistry was probably not the best idea. I'm not saying he needs to completely popcap it up with fuzzy cute critters and glowing lights and music, but there's got to be a good middle ground there. Try to ease people in with a nice environment (not to mention a friendlier tutorial) and lead them through the easy puzzles for a while first. Also he's quite right that tying the story to the ability to finish the entire game was no good. I was glad to see that they actually posted the entire story on their website, so I finally got to see how it turns out.
Of course there's plenty of room for the brain busting puzzles that take days to solve, but maybe not in the main campaign. Build that around concepts that can be completed by the average person who likes puzzle games, instead of the average person who likes differential calculus. I still think it was a brilliant game, but I freely admit to getting stuck in the middle and giving up. I was left wanting more, but my only options were puzzles that I inevitably gave up on after staring at for half an hour and making no progress.
He's been making these sorts of games for a while, though, and previously they were pretty much limited to the more hardcore fans. Now that he's had some breakout success with a wider audience, and clearly learned some important things from it, I look forward to whatever his next game might be with great interest.