[ajmoir]

That would be a good thing.

The only issue I see is that YC seems to be entirely focused on success in the Web 2.0 and software arenas. PARC had a much broader remit.

It appears to me that nobody is doing blue sky research in the computing field. Rather, it has all become a case of building a better mouse trap.

[pg]

I didn't mean we'd literally do the same things as Parc. I meant more that YC might be one of those cases where, in retrospect, people were surprised that so many different things came from one place. The phenomenon isn't limited to technology. The New Yorker in the mid 20th century was another.

[nostrademons]

"It appears to me that nobody is doing blue sky research in the computing field. Rather, it has all become a case of building a better mouse trap."

You don't hang around the Haskell crowd very much. ;-)

Seriously, blue-sky research is alive and well in the computing field. You just don't hear about it much because blue-sky research, by definition, is useless. And usually highly technical as well. Techcrunch gets far more readers by covering the latest Reddit clone than Lambda: The Ultimate gets by covering the latest advances in dependent typing.

[ajmoir]

I follow LtU not always able to comprehend the points made after all I program in a language worse than blub, I kid you not.

But nothing on there has ever come across as radical more of an evolution and sometimes it's just plain ego stroking.

As for Haskell, it's in danger of losing it's chance for wide spread acceptance. The further they take it away from being a good language for multi-processor machines the further away it is from being useful.

With the death of Von Neumann in sight, at last, Haskell is galloping away from it's biggest selling point to the average pgmr. I'm starting to think Erlang might become a mainstream language instead of Haskell. Would sort of prove the need for a language to have a single designer.

The sort of things I'd hoped to see in research are :- Genetic design of machines with 1000+ simple cpus with multi level communication and storage. Proof that the brain works on a quantum level. Cyc tied into machine learning so that it auto updates it's rules. Rules for understanding complexity from simple components.

Basically, I want to see the big problems that are just within grasp tackled.