Anyone who knows how to program in Lisp is smart.
occamsoftware.blogspot.com
occamsoftware.blogspot.com
Lisp is a language. A language is a tool. If you can use a tool to build great things, you are good. Perhaps your tool is not for the weak-minded, but wielding a tool is not enough to say anything. A three section staff requires godly dexterity, but is inferior to the longsword in combat. The latter is easier to pick up and thrash around, but way more difficult to master.
I feel like saying this because anybody who programs is likely to be able to pick up a new language and play around with it; the premises are the same. Anybody who speaks can likely pick up a new language and make sentences. You might find, say, the subject-object-verb ordering particularly intersting and conclude anyone who speaks Japanese is smart -- but it's more about practice and application.
If you decompile your mind it's probably half spaghetti code. If you decompile Feynman it'd probably be strangely productive infinite loops of sorts.
However, if you take someone who programs in Basic, and try to teach them Lisp, it will take them a long time, because there are lots of new concepts to learn. Similarly, if you take an English speaker and try to teach them Chinese, there will be a lot of stuff to learn, and it will take a long time.
An English speaker who has mastered Japanese will have gone from an SVO (subject verb object) language to an SOV language (Chinese is not as convenient an example, because it is also SVO). This is a completely different "paradigm of thinking," just for the comparison. Your brain needs to sort the order of concepts before your initiate your utterance. You need to match particles to their respective concepts. For anybody who never learned such a system, it doesn't come easily. But those who have gone to a sufficient level will become more perceptive of the nuances and particulars of "language" as a whole. This knowledge is far less accessible to those who only know one language. A linguist who only speaks one language will not be taken seriously.
Similarly, a competent programmer is likely to be versed in several languages, and likely they will fall into different paradigms. A programmer who only does LISP is likely to not have the same level of knowledge about the computer system than one who knows several languages. After all, you are looking at multiple ways of generating 1s and 0s. As such, knowing LISP certainly cannot imply anything. Of course, the author anticipated this response, and added a loophole in the article, which I consider a bit too wide, hence the criticism.
There are visible reasons why LISP trains your thinking. My post above is pretty poorly written, but the reference to spaghetti code is like the BASIC example you mention. Supposedly, good programming practice discourages spaghetti code, and BASIC was taken as the bad example. But people think in spaghetti code. If you pry open von Neumann's brain, it would probably be uber-spaghetti code. It would probably decompile into something more BASIC-like than LISP-like, but he's insanely smart.
Yes, it's the logic training. But if that was the theme, the article should have addressed it clearly. I just felt like pointing out how the current topic in the blog does a rather weak generalization, which helps more to propagate misconceptions than it does to educate.
With regards to my constraint "(and I mean really program)", I merely wanted to exclude those people who claim to "know about" LISP, or who are not competent in LISP (and I think I have a weaker threshold for competent than you).
I guess I tend to disagree with your statement:
"Similarly, a competent programmer is likely to be versed in several languages, and likely they will fall into different paradigms."
I think "a competent programmer" is a low bar (70 percential?). I don't think it's true that competent programmers are likely to be "competent" in several languages that fall into different paradigms.
Also, I feel like LISP's paradigm is so far removed from most others (and with arguably a superset of concepts), that any programmer who is competent in LISP IS likely to be competent is several other languages that span different paradigms. Agreed if there exists one programmer out there who ONLY knows LISP, they lack the complete conceptual understanding I'm asserting.
It seems to me that programmers who tend to write spaghetti code and/or over complicate matters will abandon LISP out of fustration before they achieve "competency in LISP".
The point of the article was to assert that "competency in LISP" implies much more than "competency in LISP" and that this statement is not true of most other programming languages.
Sorry you feel this article propagated misconceptions, I'll try to do better next time.
Great to hear feedback though, and you found a weakness in my argument. My preconceptions of a "competent programmer" are often based on qualities I wish I had :)
Lisp's paradigm is sufficiently different from the normal imperative languages, e.g. C or Python, that grokking Lisp means you understand more than one paradigm. It's this that's valuable, not knowledge of Lisp itself.
You often find that coders who know multiple paradigms will write interpreters for little languages in the implementation language available to them in order to better represent the problem, rather than attempting a dumb implementation in the bare implementation language.
There used to be a good book by Sam Kamin, Programming Languages: An Interpreter-Based Approach. It examines cut-down versions of different languages, e.g. Lisp and Prolog, implementing an interpreter for each of them in an imperative language to help explain how they work. Norman Ramsey was working on an update but I don't know if it's been published yet. Some universities seem to have access to a pre-release: http://www.cs.uml.edu/~dimock/courses/languages/Spring2005/mainpage.html (Disclaimer: I provided review comments to Ramsey for a couple of chapters back in 2002.)
Where Lisp is useful, is that just about any paradigm and his dog can be implemented in Lisp.
This was exactly my point... grokking Lisp is a hallmark of intelligence. I don't think there is any other programming language that rises to this level.
I think the second paragraph of his Python article captures much of my intent. But still, there seems to be something just a bit different about LISP than most other langauges.
Perhaps it is LISP's durribility or the fact that over time different (smart) people develop new lanuages and eventually begin to incorporate more and more features of LISP. LISP clearly got some things right in a fundamental way that every one seems to be rediscovering every decade. This leads me to believe that if you understand LISP, you have a deeper understanding of how things really fit together.
With specific regards to your question, I'd include most main stream programming languages (but that's kinda obvious). At this point, I'd include Ruby in this list, but may not Python. Although even in the case of Python, it strikes me as being very similar to other OO languages (better perhaps, but similar none-the-less).
Seriously... When I used Lisp in college I remember thinking that it felt like cheating. It felt like using a graphing calculator to solve polynomials in high school. The only other languages I'd used in class were C, C++, Java, and 68K assembler, and it was hard to shake the idea that serious programming was supposed to be painful and low-level.
But I can tell one thing from my experience: All competent Lisp programmers I know are also pretty good at C, C++, Java, or even Assembly. Even if they don't initially, they quickly become good. So, in empirical sense, using Lisp as one filtering factor can work. (But be aware: The fact that they can write good Java code doesn't mean they want to do so).
The more valid comparison might be "Anyone who can speak Latin must be smart," and that is probably not far wrong.
Because if you simply "know" a language, it means nothing unless you quantify some experience behind that simple awareness of the language syntax.
A lot of people say they "know", when they mean "know about."