“When people who can’t think logically design large systems, those systems become incomprehensible. And we start thinking of them as biological systems. And since biological systems are too complex to understand, it seems perfectly natural that computer programs should be too complex to understand.”
Simultaneously a straw man and a false dichotomy. Not written by someone who understands logic?
What he may have meant is that programmers using the biological approach with things like information hiding, guard functions, and testing built complex programs that usually work as intended. That’s without knowing anything about formal logic or mathematical aspects. Writers covering things like LISP used to compare it to biological approaches as arguments it was more adaptable whereas the formalized stuff failed do to rigidity and slow-moving. Just reading Leslie’s remark, someone might assume all biologically-inspired approaches were barely comprehensible or failures whereas the formal or logical methods stayed outperforming them. Most of the latter actually failed.
I still enjoyed reading it despite that inaccuracy. Leslie’s mind is interesting to watch in action with down-to-earth style. This reminded me of a computer scientist who thought like a biologist to overcome limitations CompSci folks were facing. Led him to do everything from invent massively-parallel processing to using evolution to try to outperform human designers. Always claimed biology was better. A lot of better write-ups are paywalled or disappearing with Old Web but I can try to dig some out this week if you’re interested.
With the way ML/AI is going, it’s quite possible many future systems could be much closer to biology than human design. An AI system-design software will just do whatever works as long as its optimization function says it’s good.
I am in no way questioning Lamport’s brilliance nor contributions in general. However most people, brilliant or otherwise, have blind spots. I believe he’s betrayed some of his here, and that in itself is interesting and worth reading.
“When people who can’t think logically design large systems, those systems become incomprehensible. And we start thinking of them as biological systems. And since biological systems are too complex to understand, it seems perfectly natural that computer programs should be too complex to understand.”
Simultaneously a straw man and a false dichotomy. Not written by someone who understands logic?
The author is Leslie Lamport, who won the 2013 Turing Award for his work on distributed algorithms.
I’m aware of that. My question is rhetorical.
What he may have meant is that programmers using the biological approach with things like information hiding, guard functions, and testing built complex programs that usually work as intended. That’s without knowing anything about formal logic or mathematical aspects. Writers covering things like LISP used to compare it to biological approaches as arguments it was more adaptable whereas the formalized stuff failed do to rigidity and slow-moving. Just reading Leslie’s remark, someone might assume all biologically-inspired approaches were barely comprehensible or failures whereas the formal or logical methods stayed outperforming them. Most of the latter actually failed.
I still enjoyed reading it despite that inaccuracy. Leslie’s mind is interesting to watch in action with down-to-earth style. This reminded me of a computer scientist who thought like a biologist to overcome limitations CompSci folks were facing. Led him to do everything from invent massively-parallel processing to using evolution to try to outperform human designers. Always claimed biology was better. A lot of better write-ups are paywalled or disappearing with Old Web but I can try to dig some out this week if you’re interested.
Please do dig it up, I’m quite intrigued to see where their solutions worked well, and where they didn’t.
With the way ML/AI is going, it’s quite possible many future systems could be much closer to biology than human design. An AI system-design software will just do whatever works as long as its optimization function says it’s good.
I am in no way questioning Lamport’s brilliance nor contributions in general. However most people, brilliant or otherwise, have blind spots. I believe he’s betrayed some of his here, and that in itself is interesting and worth reading.