It's not much of a slam when it's completely true. I've practiced Geophysics, aerospace engineering and computer science, and I can tell you that, in general, CS lacks all the appropriate education of reliability, durability, precision, customer responsibility, deference to reality, notational consistency*, thoroughness of documentation, and subjugation of personal pride to achieving mission objectives that are absolutely essential in a real science or engineering. The folks to whom I've spoken who say that CS is not less rigorous simply don't have the experience necessary to comment, or else they work on financial software. I've been fortunate to work with some of the best in the CS industry, but... aerospace is still an order of magnitude more difficult and demanding.
But it's clear you don't understand my point. So, to clarify:
I'm not saying software can't or hasn't make major contributions. I'm just saying compared to real engineerings and sciences, CS (and education in it) is unreliable, flimsy, poorly documented, and lacks proper customer-demand and mission-objective focus.
Apple and Microsoft's growing recent failures in each passing update only serve to prove my point. I only partially blame the tool sets. As iOS developers, we still don't even have tool sets which can search our code for syntax-context aware expressions, like "find all calls to a method named 'foo()' on instances of the 'bar' class". We're left with
text searches and reg-ex searches which ask me to construct a syntax so dense and fathomless that the task is usually impractical. Performance profiling tools won't even tell me what the arguments were to function calls which take the most time. If I had run my 10,000 tests for my aerospace dissertation research and failed to match each test result with the data which went in to it, I would have been laughed out of the department, but that's the state of the art in CS. If there is any tendency to improve these failures in the CS industry, excellent. I kind of think maybe Chris Lattner and his team will deliver something incredibly useful like this at some point, but that might just be my little man-crush on him for bring us Obj-C-ARC.
And as for documentation, if a man-page or non-free Apple developer's guide even approached the kind of detail I get from a free data sheet from Digi-Key when browsing for electronic components, I would be shocked. The "header" style documentation I've seen is worse than RadioShack's 74xx series data sheets which sometimes wouldn't tell you which pin was which pin number. In fact, that might be a fair comparison, Radio Shack documentation is to computer science as digi-key's free data sheets are to rigorous engineerings. Apple's documentation on in-app purchase is about as good as it gets, and that amounts to handing me a sample implementation circuit.
And if you manage to prove me wrong, well then I'll just be happy, because things will be better than I think they are.
* Notational consistency as applied to science. Engineers are terrible at notational consistency, even in a single industry.
