To fully explain, it would require a blog post, and it has been covered before, but i'll make a try:
- Proper stack traces. Debugging Common Lisp code is not only easy, it is far easier than in 90% of programming platforms out there. Clojure is severly lacking here; it has been promised that there will be an improvement; i will be watching since this will greatly improve Clojure
- "Lisp-2 versus Lisp-1": On Common Lisp, functions and variables have separate namespaces so they don't conflict. This makes writing code, especially macros much easier. Macros are one of the biggest reasons to use a Lisp dialect in the first place, so this is a rather strong point.
- Object orientation facilities: CLOS, CL Object System, is arguably the most powerful OOP system on a mainstream language. You should really take a look at it, it can just blow your mind. Clojure has no comparable OOP implementation. I recommend CLOS to anyone that feels disillusioned with OOP.
I could stop here, because the three points above are very strong differences, but there are some other points to consider:
- Standarization and portability (1/2): CL is an ANSI standard, and there are many ANSI-compliant, mature, compilers: LispWorks, SBCL, Allegro CL, ABCL, Clozure CL, ECL, CLISP, and others. There are features that the ANSI standard does not cover, like interacing with C libraries, but there are already portable libraries that allow you to also do this on a portable way. So, in this way, your code can be compiled by those implementations, often without any change needed at all.
- Standarization and portability (2/2): The ABCL compiler allows you to target the JVM and call Java code from Lisp or call Lisp code from Java. The Clozure compiler lets you target the LLVM. Practically most modern platforms can be compiled from CL to native code by using the suitable compiler.
- Interfacing: As mentioned above, you can interface with Java, and with C, as needed. C interfacing is very easy, btw.
- Speed: CL can be very fast. For example SBCL is really fast, particularly with numbers where it can be as fast as C. In general it is in the same speed of Java code running under the Oracle's latest JVM. I think that CL code under SBCL probably should be much faster than Clojure running under JVM.
- Number data types: CL supports, with high perfomance and natively: Complex numbers, Rational (fractional) numbers, floating numbers (with the IEEE standard), integers, binary numbers, and arbitrary-length numbers. Bonus: The same typical operators (+,-,*,/) work perfectly with them, so using them is simple. In this feature, CL fares even better than Fortran (seriously.)
- Static typing? CL is not statically typed, but the standard allows you to declare the type of your variables. A compiler like SBCL will catch many type errors during compilation (and of course they will also be catched at runtime). Doing this has also the big bonus of increasing performance dramatically.
- Mainstream: While not popular, Common Lisp is a mainstream language, so there are tons of books and resources out there, as well as having been used successfully complex production systems, like for example auto-piloting the Deep Space 1 spaceship (NASA) for days.
- Proper stack traces. Debugging Common Lisp code is not only easy, it is far easier than in 90% of programming platforms out there. Clojure is severly lacking here; it has been promised that there will be an improvement; i will be watching since this will greatly improve Clojure
- "Lisp-2 versus Lisp-1": On Common Lisp, functions and variables have separate namespaces so they don't conflict. This makes writing code, especially macros much easier. Macros are one of the biggest reasons to use a Lisp dialect in the first place, so this is a rather strong point.
- Object orientation facilities: CLOS, CL Object System, is arguably the most powerful OOP system on a mainstream language. You should really take a look at it, it can just blow your mind. Clojure has no comparable OOP implementation. I recommend CLOS to anyone that feels disillusioned with OOP.
I could stop here, because the three points above are very strong differences, but there are some other points to consider:
- Standarization and portability (1/2): CL is an ANSI standard, and there are many ANSI-compliant, mature, compilers: LispWorks, SBCL, Allegro CL, ABCL, Clozure CL, ECL, CLISP, and others. There are features that the ANSI standard does not cover, like interacing with C libraries, but there are already portable libraries that allow you to also do this on a portable way. So, in this way, your code can be compiled by those implementations, often without any change needed at all.
- Standarization and portability (2/2): The ABCL compiler allows you to target the JVM and call Java code from Lisp or call Lisp code from Java. The Clozure compiler lets you target the LLVM. Practically most modern platforms can be compiled from CL to native code by using the suitable compiler.
- Interfacing: As mentioned above, you can interface with Java, and with C, as needed. C interfacing is very easy, btw.
- Speed: CL can be very fast. For example SBCL is really fast, particularly with numbers where it can be as fast as C. In general it is in the same speed of Java code running under the Oracle's latest JVM. I think that CL code under SBCL probably should be much faster than Clojure running under JVM.
- Number data types: CL supports, with high perfomance and natively: Complex numbers, Rational (fractional) numbers, floating numbers (with the IEEE standard), integers, binary numbers, and arbitrary-length numbers. Bonus: The same typical operators (+,-,*,/) work perfectly with them, so using them is simple. In this feature, CL fares even better than Fortran (seriously.)
- Static typing? CL is not statically typed, but the standard allows you to declare the type of your variables. A compiler like SBCL will catch many type errors during compilation (and of course they will also be catched at runtime). Doing this has also the big bonus of increasing performance dramatically.
- Mainstream: While not popular, Common Lisp is a mainstream language, so there are tons of books and resources out there, as well as having been used successfully complex production systems, like for example auto-piloting the Deep Space 1 spaceship (NASA) for days.