Related
I am implementing an interpreter that codegen to another language using Racket. As a novice I'm trying to avoid macros to the extent that I can ;) Hence I came up with the following "interpreter":
(define op (open-output-bytes))
(define (interpret arg)
(define r
(syntax-case arg (if)
[(if a b) #'(fprintf op "if (~a) {~a}" a b)]))
; other cases here
(eval r))
This looks a bit clumsy to me. Is there a "best practice" for doing this? Am I doing a totally crazy thing here?
Short answer: yes, this is a reasonable thing to do. The way in which you do it is going to depend a lot on the specifics of your situation, though.
You're absolutely right to observe that generating programs as strings is an error-prone and fragile way to do it. Avoiding this, though, requires being able to express the target language at a higher level, in order to circumvent that language's parser.
Again, it really has a lot to do with the language that you're targeting, and how good a job you want to do. I've hacked together things like this for generating Python myself, in a situation where I knew I didn't have time to do things right.
EDIT: oh, you're doing Python too? Bleah! :)
You have a number of different choices. Your cleanest choice is to generate a representation of Python AST nodes, so you can either inject them directly or use existing serialization. You're going to ask me whether there are libraries for this, and ... I fergits. I do believe that the current Python architecture includes ... okay, yes, I went and looked, and you're in good shape. Python's "Parser" module generates ASTs, and it looks like the AST module can be constructed directly.
https://docs.python.org/3/library/ast.html#module-ast
I'm guessing your cleanest path would be to generate JSON that represents these AST modules, then write a Python stub that translates these to Python ASTs.
All of this assumes that you want to take the high road; there's a broad spectrum of in-between approaches involving simple generalizations of python syntax (e.g.: oh, it looks like this kind of statement has a colon followed by an indented block of code, etc.).
If your source language shares syntax with Racket, then use read-syntax to produce a syntax-object representing the input program. Then use recursive descent using syntax-case or syntax-parse to discern between the various constructs.
Instead of printing directly to an output port, I recommend building a tree of elements (strings, numbers, symbols etc). The last step is then to print all the elements of the tree. Representing the output using a tree is very flexible and allows you to handle sub expressions out of order. It also allows you to efficiently concatenate output from different sources.
Macros are not needed.
I'm trying to learn Common Lisp, and found something unexpected (to me) when trying something out in the repl. Based on order of execution in most programming languages, and the great first class function support I'd always heard about from lisp, I'd think the following should work:
((if t 'format) t "test")
In Ruby I can do:
if true
Object.method(:puts)
end.call("test")
My thinking in how the above lisp code should work is that it should evaluate the inner lisp form, return format, then begin evaluating the outer lisp form, with format then being the first atom. I'd read that the first form needs to be a symbol, so I also tried ((if t format) t "test") even though my initial thought was that this would try to evaluate format before returning from the inner form.
I've noticed that sometimes lisp forms need to be preceded by #' in order for their results to be callable, but using (#'(if t 'format) t "test") doesn't work either. I'm sure I'm just misunderstanding something basic as I'm pretty new to lisp, but what's going on here?
Common Lisp doesn't evaluate the first element of an expression normally. It has to be either a literal symbol naming a function, or a lambda expression.
If you want to call a function determined dynamically, you need to use the FUNCALL function:
(funcall (if t 'format) t "test")
This is analogous to the need to use the .call() method in Ruby.
What you tried would work in some other Lisp dialects, such as Scheme.
I would like to take Emacs Lisp code that has been macro expanded and unmacro expand it. I have asked this on the Emacs forum with no success. See:
https://emacs.stackexchange.com/questions/35913/program-rewriting-systems-unexpanded-a-defmacro-given-a-list-of-macros-to-undo
However one would think that this kind of thing, S-expression transformation, is right up Lisp's alley. And defmacro is I believe available in Lisp as it is in Emacs Lisp.
So surely there are program transformation systems, or term-rewriting systems that can be adapted here.
Ideally, in certain situations such a tool would be able to work directly off the defmacro to do its pattern find and replace on. However even if I have to come up with specific search and replace patterns manually to add to the transformation system, having such a framework to work in would still be useful
Summary of results so far: Although there have been a few answers that explore interesting possibilities, right now there is nothing definitive. So I think best to leave this open. I'll summarize some of the suggestions. (I've upvoted all the answers that were in fact answers instead of commentary on the difficulty.)
First, many people suggest considered the special form of macros that do expansion only,or as Drew puts it:
macro-expansion (i.e., not expansion followed by Lisp evaluation).
Macro-expansion is another way of saying reduction semantics, or
rewriting.
The current front-runner to my mind is in phils post where he uses a pattern-matching facility that seems specific to Emacs: pcase. I will be exploring this and will post results of my findings. If anyone else has thoughts on this please chime in.
Drew wrote a program called FTOC whose purpose was to convert Franz Lisp to Common Lisp; googling turns up a comp.lang.lisp posting
I found a Common Lisp package called optima with fare-quasiquote. Paulo thinks however this might not be powerful enough since it doesn't handle backtracking out of the box, but might be programmed in by hand. Although the generality of backtracking might be nice, I'm not convinced I need that for the most-used situations.)
Side note: Some seem put off by the specific application causing my initial interest. (But note that in research, it is not uncommon for good solutions to get applied in ways not initially envisioned.)
So in that spirit, here are a couple of suggestions for changing the end application. A good solution for these would probably translate to a solution for Emacs Lisp. (And if if helps you to pretend I'm not interested in Emacs Lisp, that's okay with me). Instead of a decompiler for Emacs Lisp, suppose I want to write a decompiler for clojure or some Common Lisp system. Or as suggested by Sylwester's answer, suppose I would like to automatically refactor my code by taking into account the benefit of using more concise macros that exist or that have gotten improved. Recall that at one time Emacs Lisp didn't have "when" or "unless" macros.
30-some years ago I did something similar, using macrolet.
(Actually, I used defmacro because we had only an early implementation of Common Lisp, which did not yet have macrolet. But macrolet is the right thing to use.)
I didn't translate macro-expanded code to what it was expanded from, but the idea is pretty much the same. You will come across some different difficulties, I expect, since your translation is even farther away from one-to-one.
I wrote a translator from (what was then) Franz Lisp to Common Lisp, to help with porting lots of existing code to a Lisp+Prolog-machine project. Franz Lisp back then was only dynamically scoped, while Common Lisp is (in general) lexically scoped.
And yes, obviously there is no general way to automatically translate Lisp code (in particular), especially considering that it can generate and then evaluate other code - but even ignoring that special case. Many functions are quite similar, but there is the lexical/dynamic difference, as well as significant differences in the semantics of some seemingly similar functions.
All of that has to be understood and taken for granted from the outset, by anyone wanting to make use of the results of translation.
Still, much that is useful can be done. And if the resulting code is self-documenting, telling you what it was derived from etc., then when in the resulting context you can decide just what to do with this or that bit that might be tricky (e.g., rewrite it manually, from scratch or just tweak it). In practice, lots of code was easily converted from Franz to Common - it saved much reprogramming effort.
The translator program was written in Common Lisp. It could be used interactively as well as in batch. When used interactively it provided, in effect, a Franz Lisp interpreter on top of Common Lisp.
The program used only macro-expansion (i.e., not expansion followed by Lisp evaluation). Macro-expansion is another way of saying reduction semantics, or rewriting.
Input Franz-Lisp code was macro-expanded via function-definition mapping macros to produce Common-Lisp code. Code that was problematic for translation was flagged (in code) with a description/analysis that described the situation.
The program was called FTOC. I think you can still find it, or at least references to it, by googling (ftoc lisp). (It was the first Lisp program I wrote, and I still have fond memories of the experience. It was a good way to learn both Lisp dialects and to learn Lisp in general.)
Have fun!
In general, I don't think you can do this. The expansion of an lisp macro is Turing complete, so you have to be able to predict the output of a program which could have arbitrary input.
There are some simple things that you could do. defmacros with backquoted forms in appear fairly similar in the output form and might be detected. This sort of heuristic would probably get you a long way.
What I don't understand is your use case. The macro-expanded version of a piece of code is usually only present in the compiled (or in emacs-lisp byte-compiled) form.
Ok so other people have pointed out the fact that this problem is impossible in general. There are two hard parts to this problem: one is that it could be a lot of work to find a preimage of some code fragment through a macro and it is also impossible to determine whether a macro was called or not—there are examples where one may write code which could have come from a macro without using that macro. Imagine for the sake of illustration an sha macro which expands to the SHA hash of the string literal passed to it. Then if you see some sha hash in your expanded code, it would obviously be silly to try to unexpand it. But it may be that the hash was put into the code as a literal, e.g. referencing a specific point in the history of a git repository so it would also be unhelpful to unexpand the macro.
Tractable subproblems
Let me preface this by saying that whilst these may be a little tractable, I still wouldn’t try to solve this problem.
Let’s ignore all the macros that do weird things (like the example above) and all the macros that are just as likely to not have been used in the original (e.g. cond vs if) and all the macros which generate complex code which seems like it would be difficult to unravel (e.g. loop, do, and backquote. Annoyingly these difficult cases are some of those which you would perhaps most want to unexpand). The type this leaves us with (that I’d like to focus on) are macros which basically just reduce boilerplate, e.g. save-excursion or with-XXXX. These are macros whose implementation consists of possibly making some fresh symbols (via gensym) and then having a big simple backquoted block of code. I still think it would be too hard to automatically go from defmacro to a function for unexpansion but I think you could attack some of these on a case-by-case basis. Do this by looking for the forms generated by the macro that delimit (I.e. begin/end) the expanded code. I can’t really offer much beyond that. This is still a hard problem and I don’t think any existing solutions (to other problems) will get you very far on your way.
A further complication I understand is that you do not start at the macroexpanded code but rather at the bytecode. Without knowing anything about the elisp compiler, I worry that more information would be lost in the compilation step and you would have to undo that as well, e.g. perhaps it is hard to determine which code goes inside a let or even when a let begins, or bytecode starts using goto type features even though elisp doesn’t have them.
You suggest that the reason you would like to unexpand macros is so you can decompile bytecode which sometimes comes up in the Emacs debugger and that this would be useful as even though the source code is available in theory, it isn’t always at your fingertips. I put it to you that if you want to make your life debugging elisp easier it would be more worthwhile to figure out how to have the Emacs debugger always take you to the source code for internal functions. This might involve installing extra debugging related packages or downloading the Emacs source code and setting some variable so Emacs knows where to find it or compiling Emacs yourself from source. I don’t really know about that but I bet getting thrown into bytecode instead of source would have been enough of a problem for Emacs developers over the past thirty years that a solution to that problem does exist.
If however what you really want to do is to try to implement a decompiler for elisp then I suppose that’s what you should do. A final observation is that while Lisp provides facilities which make manipulating Lisp code easy, this doesn’t help much with decompiling as all these facilities can be used in compilation so there are infinitely more patterns one might want to detect than in e.g. a C decompiler. Perhaps scheme style macros would be easier to unexpand, although they would still be hard.
If you’re decompiling because you want to give a better idea of which exact subexpression rather than line is being evaluated (normally Lisp debuggers work on expressions not lines anyway) in the debugger then perhaps it would actually be useful to see the code at the expanded level rather than the unexpanded one. Or perhaps it would be best to see both and maybe in between as well. Keeping track of what’s what through forwards macroexpansion is already difficult and fiddly. Doing it in reverse certainly won’t be easier. Good luck!
Edit: seeing as your not currently using Lisp anyway, I wonder if you might have more success using something like prolog for your unexpanding. You’d still have to manually write rules but I think it would be a large amount of work to try to derive rules from macro definitions.
I would like to take Emacs Lisp code that has been macro expanded and unmacro expand it.
Macros generate arbitrary expressions, which may contain macros recursively. You have no general way to revert the transformations, because it's not pattern-based.
Even if macros were pattern-based, they could still be infinite.
Even if macros were not infinite, they can certainly contain bugs in expansions of patterns that never matched. Given arbitrary code to try to unwind, it could match an expansion that looks like the code and try to revert to its pattern. Without bugs, you could still abuse this.
Even if you could revert macro expansion, some macros expand to the same code. An approach could be signalling a warning with a restart when all reversions expand equally minus the operator, such that if the restart doesn't handle the signal, it would choose the first expansion; and otherwise signalling an error with a restart, such that if the restart doesn't handle the signal, it errors. Or you could configure it to choose certain macros under certain conditions, such as in which package the code was found.
In practice, there are very few cases where reverting an expansion makes any sense. It could be a useful development tool that suggests macros, but I wouldn't generally rely on it for whole source transformations.
One way you could achieve what you want is through a controlled pattern matching. You could initially create patterns manually, which would already handle cases you care about directly, such as the ones you mention:
(if (not <cond>) <expr>) and (if (not <cond>) (progn <&expr>)) to (unless <cond> <&expr>)
You'd have to decide whether null would be equivalent to not. I personally don't mix the boolean meaning of nil with that of empty list or something else, e.g. no result, nothing found, null object, a designator, etc. But perhaps Lisp code as old as that in Emacs just uses them interchangeably.
(if <cond> <expr>) and (if <cond> (progn <&expr>)) to (when <cond> <&expr>)
If you feel like improving code overall, include cond with a single condition. And be careful with cond clauses with only the condition.
You should have a few dozen more, to see how the pattern matching behaves with more patterns to match in terms of time (CPU) and space (memory).
From the description of fare-quasiquote, optima doesn't support backtracking, which you probably want.
But you can do backtracking with optima by yourself, using recursion on complex inner patterns, and if nothing matches, return a control value to keep searching for matching patterns from the outer input.
Another approach is to treat a pattern as a description of a state machine, and handle each new token to advance the current state machines until one of them reaches the end, discarding the state machines that couldn't advance. This approach may consume more memory, depending on the amount of patterns, the similarity between patterns (if many have the same starting token, many state machines will be generated on a matching token), the length of the patterns and, last but not least, the length of the input (s-expression).
An advantage of this approach is that you can use it interactively to see which patterns have matched the most tokens, and you can give weights to patterns instead of just taking the first that matches.
A disadvantage is that, most probably, you'll have to spend effort to develop it.
EDIT: I just lousily described a kind of trie or radix tree.
Once you got something working, maybe try to obtain patterns automatically. This is really hard, you must probably limit it to simple backquoting and accept the fact you can't generalize for anything that contains more complex code.
I believe the hardest will be code walking, which is hard enough with source code, but much more with macro-expanded code. Perhaps if you could expand the whole picture a bit further to understand the goal, maybe someone could suggest a better approach other than operating on macro-expanded code.
However one would think that this kind of thing, S-expression transformation, is right up Lisp's alley. And defmacro is I believe available in Lisp as it is in Emacs Lisp.
So surely there are program transformation systems, or term-rewriting systems that can be adapted here.
There's a huge step from expanding code with defmacro and all that generality. Most Lisp developers will know about hygienic macros, at least in terms of symbols as variables.
But there's still hygienic macros in terms of symbols as operators1, code walking, interaction with a containing macro (usually using macrolet), etc. It's way too complex.
1.
Common Lisp evaluates the operator in a compound form in the lexical environment, and probably everyone makes macros that assume that the global macro or function definition of a symbol will be used.
But it might not be so:
(defmacro my-macro-1 ()
`1)
(defmacro my-macro-2 ()
`(my-function (my-macro-1)))
(defun my-function (n)
(* n 100))
(macrolet ((my-macro-1 ()
`2))
(flet ((my-function (n)
(* n 1000)))
(my-macro-2)))
That last line will expand to (my-function (my-macro-2)), which will be recursively expanded to (my-function 2). When evaluated, it will yield 2000.
For proper operator hygiene, you'd have to do something like this:
(defmacro my-macro-2 ()
;; capture global bindings of my-macro-1 and my-function-1 by name
(flet ((my-macro-1-global (form env)
(funcall (macro-function 'my-macro-1) form env))
(my-function-global (&rest args)
;; hope the compiler can optimize this
(apply 'my-function args)))
;; store them globally in uninterned symbols
;; hopefully, no one will mess with them
(let ((my-macro-1-symbol (gensym (symbol-name 'my-macro-1)))
(my-function-symbol (gensym (symbol-name 'my-function))))
(setf (macro-function my-macro-1-symbol) #'my-macro-1-global)
(setf (symbol-function my-function-symbol) #'my-function-global)
`(,my-function-symbol (,my-macro-1-symbol)))))
With this definition, the example will yield 100.
Common Lisp has some restrictions to avoid this, but it only states the consequences are undefined when (re)defining symbols in the common-lisp package, globally or locally. It doesn't require errors or warnings to be signaled.
I don't think it is possible to do this in general, but you can undo a pattern back into a macro use for every match if you supply code for each unmacroing. Code that mixed cond and if will end up being just if and your code would remove all if into cond making the reverse not the same as the starting point. The more macros you have and the more they expand into each other the more uncertain of the end result will be of the starting point.
You could have rules such that if is not translated into cond unless you used one of the features, like more than one predicate or implicit progn, but you have no idea if the coder actually did use cond everywhere because he liked in consistent regardless. Thus your unmacroing will acyually be more of a simplification.
I don't believe there's a general solution to that, and you certainly
can't guarantee that the structure of the output would match that of
the original code, and I'm not going near the idea of auto-generating
patterns and desired transformations from macro definitions; but you
might achieve a simple version of this with Emacs' own pcase pattern
matching facility.
Here's the simplest example I could think of:
With reference to the definition of when:
(defmacro when (cond &rest body)
(list 'if cond (cons 'progn body)))
We can transform code using a pcase pattern like so:
(let ((form '(if (and foo bar baz) (progn do (all the) things))))
(pcase form
(`(if ,cond (progn . ,body))
`(when ,cond ,#body))
(_ form)))
=> (when (and foo bar baz) do (all the) things)
Obviously if the macro definitions change, then your patterns will
cease to work (but that's a pretty safe kind of failure).
Caveat: This is the first time I've written a pcase form, and I
don't know what I don't know. It seems to work as intended, though.
In Lisp, any program's code is actually a valid data structure. For example, this adds one and two together, but it's also a list of three items.
(+ 1 2)
What benefit does that provide? What does that enable you to do that's impossible and/or less elegant in other languages?
To make things a little clearer with respect to code representation, consider that in every language code is data: all you need is strings. (And perhaps a few file operations.) Contemplating how that helps you to mimic the Lisp benefits of having a macro system is a good way for enlightenment. Even better if you try to implement such a macro system. You'll run into the advantages of having a structured representation vs the flatness of strings, the need to run the transformations first and define "syntactic hooks" to tell you where to apply them, etc etc.
But the main thing that you'll see in all of this is that macros are essentially a convenient facility for compiler hooks -- ones that are hooked on newly created keywords. As such, the only thing that is really needed is some way to have user code interact with compiler code. Flat strings are one way to do it, but they provide so little information that the macro writer is left with the task of implementing a parser from scratch. On the other hand, you could expose some internal compiler structure like the pre-parsed AST trees, but those tend to expose too much information to be convenient and it means that the compiler needs to somehow be able to parse the new syntactic extension that you intend to implement. S-expressions are a nice solution to the latter: the compiler can parse anything since the syntax is uniform. They're also a solution to the former, since they're simple structures with rich support by the language for taking them apart and re-combining them in new ways.
But of course that's not the end of the story. For example, it's interesting to compare plain symbolic macros as in CL and hygienic macros as in Scheme implementations: those are usually implemented by adding more information to the represented data, which means that you can do more with those macro system. (You can also do more with CL macros since the extra information is also available, but instead of making it part of the syntax representation, it's passed as an extra environment argument to macros.)
My favorite example... In college some friends of mine were writing a compiler in Lisp. So all their data structures including the parse tree was lisp s-expressions. When it was time to implement their code generation phase, they simply executed their parse tree.
Lisp has been developed for the manipulation of symbolic data of all kinds. It turns out that one can also see any Lisp program as symbolic data. So you can apply the manipulation capabilities of Lisp to itself.
If you want to compute with programs (to create new programs, to compile programs to machine code, to translate programs from Lisp to other languages) one has basically three choices:
use strings. This gets tedious parsing and unparsing strings all the time.
use parse trees. Useful but gets complex.
use symbolic expressions like in Lisp. The programs are preparsed into familiar datastructures (lists, symbols, strings, numbers, ...) and the manipulation routines are written in the usual language provided functionality.
So what you get with Lisp? A relatively simple way to write programs that manipulate other programs. From Macros to Compilers there are many examples of this. You also get that you can embed languages into Lisp easily.
It allows you to write macros that simply transform one tree of lists to another. Simple (Scheme) example:
(define-syntax and
(syntax-rules ()
((and) #t)
((and thing) thing)
((and thing rest ...) (if thing (and rest ...) #f))))
Notice how the macro invocations are simply matched to the (and), (and thing), and (and thing rest ...) clauses (depending on the arity of the invocation), and handled appropriately.
With other languages, macros would have to deal with some kind of internal AST of the code being transformed---there wouldn't otherwise be an easy way to "see" your code in a programmatic format---and that would increase the friction of writing macros.
In Lisp programs, macros are generally used pretty frequently, precisely because of the low friction of writing them.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 8 years ago.
Improve this question
I've read The Nature of Lisp. The only thing I really got out of that was "code is data." But without defining what these terms mean and why they are usually thought to be separate, I gain no insight. My initial reaction to "code is data" is, so what?
The old fashioned view: 'it' is interactive computation with symbolic expressions.
Lisp enables easy representation of all kinds of expressions:
english sentence
(the man saw the moon)
math
(2 * x ** 3 + 4 * x ** 2 - 3 * x + 3)
rules
(<- (likes Kim ?x) (likes ?x Lee) (likes ?x Kim))
and also Lisp itself
(mapcar (function sqr) (quote (1 2 3 4 5)))
and many many many more.
Lisp now allows to write programs that compute with such expressions:
(translate (quote (the man saw the moon)) (quote german))
(solve (quote (2 * x ** 3 + 4 * x ** 2 - 3 * x + 3)) (quote (x . 3)))
(show-all (quote (<- (likes Kim ?x) (likes ?x Lee) (likes ?x Kim))))
(eval (quote (mapcar (function sqr) (quote (1 2 3 4 5)))))
Interactive means that programming is a dialog with Lisp. You enter an expression and Lisp computes the side effects (for example output) and the value.
So your programming session is like 'talking' with the Lisp system. You work with it until you get the right answers.
What are these expressions? They are sentences in some language. They are part descriptions of turbines. They are theorems describing a floating point engine of an AMD processor. They are computer algebra expressions in physics. They are descriptions of circuits. They are rules in a game. They are descriptions of behavior of actors in games. They are rules in a medical diagnosis system.
Lisp allows you to write down facts, rules, formulas as symbolic expressions. It allows you to write programs that work with these expressions. You can compute the value of a formula. But you can equally easy write programs that compute new formulas from formulas (symbolic math: integrate, derive, ...). That was Lisp designed for.
As a side effect Lisp programs are represented as such expressions too. Then there is also a Lisp program that evaluates or compiles other Lisp programs. So the very idea of Lisp, the computation with symbolic expressions, has been applied to Lisp itself. Lisp programs are symbolic expressions and the computation is a Lisp expression.
Alan Kay (of Smalltalk fame) calls the original definition of Lisp evaluation in Lisp the Maxwell's equations of programming.
Write Lisp code. The only way to really 'get' Lisp (or any language, for that matter) is to roll up your sleeves and implement some things in it. Like anything else, you can read all you want, but if you want to really get a firm grasp on what's going on, you've got to step outside the theoretical and start working with the practical.
The way you "get" any language is by trying to write some code in it.
About the "data is code" thing, in most languages there is a clear separation between the code that gets executed, and the data that is processed.
For example, the following simple C-like function:
void foo(int i){
int j;
if (i % 42 == 0){
bar(i-2);
}
for (j = 0; j < i; ++j){
baz();
}
}
the actual control flow is determined once, statically, while writing the code. The function bar isn't going to change, and the if statement at the beginning of the function isn't going to disappear. This code is not data, it can not be manipulated by the program.
All that can be manipulated is the initial value of i. And on the other hand, that value can not be executed the way code can. You can call the function foo, but you can't call the variable i. So i is data, but it is not code.
Lisp does not have this distinction. The program code is data that can be manipulated too. Your code can, at runtime, take the function foo, and perhaps add another if statement, perhaps change the condition in the for-loop, perhaps replace the call to baz with another function call. All your code is data that can be inspected and manipulated as simply as the above function can inspect and manipulate the integer i.
I would highly recommend Structure and Interpretation of Computer Programs, which actually uses scheme, but that is a dialect of lisp. It will help you "get" lisp by having you do many different exercises and goes on to show some of the ways that lisp is so usefull.
I think you have to have more empathy for compiler writers to understand how fundamental the code is data thing is. I'll admit, I've never taken a compilers course, but converting any sufficiently high-level language into machine code is a hard problem, and LISP, in many ways, resembles an intermediate step in this process. In the same way that C is "close to the metal", LISP is close to the compiler.
This worked for me:
Read "The Little Schemer". It's the shortest path to get you thinking in Lisp mode (minus the macros). As a bonus, it's relatively short/fun/inexpensive.
Find a good book/tutorial to get you started with macros. I found chapter 8 of "The Scheme
Programming Language" to be a good starting point for Scheme.
http://www.ccs.neu.edu/home/matthias/BTLS/
http://www.scheme.com/tspl3/syntax.html
By watching legendary Structure and Interpretation of Computer Programs?
In Common Lisp, "code is data" boils down to this. When you write, for example:
(add 1 2)
your Lisp system will parse that text and generate a list with three elements: the symbol ADD, and the numbers 1 and 2. So now they're data. You can do whatever you want with them, replace elements, insert other stuff, etc.
The fun part is that you can pass this data on to the compiler and, because you can manipulate these data structures using Lisp itself, this means you can write programs that write other programs. This is not as complicated as it sounds, and Lispers do it all the time using macros. So, just get a book about Lisp, and try it out.
Okay, I'm going to take a crack at this. I'm new to Lisp myself, just having arrived from the world of python. I haven't experienced that sudden moment of enlightenment that all the old Lispers talk about, but I'll tell you what I am seeing so far.
First, look at this random bit of python code:
def is_palindrome(st):
l = len(st)/2
return True if st[:l] == st[:-l-1:-1] else False
Now look at this:
"""
def is_palindrome(st):
l = len(st)/2
return True if st[:l] == st[:-l-1:-1] else False
"""
What do you, as a programmer, see? The code is identical, FYI.
If you are like me, you'll tend to think of the first as active code. It consists of a number of syntactic elements.
The second, despite its similarity, is a single syntactic item. It's a string. You interact with it as a single entity. To deal with it as code - to handle it comfortably along its syntactic boundaries - you will have to do some parsing. To execute it, you need to invoke an interpreter. It's not the same thing at all as the first.
So when we do code generation in most languages what are we dealing with? Strings. When I generate HTML or SQL with python I use python strings as the interface between the two languages. Even if I generate python with python, strings are the tool.*
Doesn't the thought of that just... make you want to dance with joy? There's always this grotesque mismatch between that which you are working with and that which you are working on. I sensed that the first time that I generated SQL with perl. Differences in escaping. Differences in formatting: think about trying to get a generated html document to look tidy. Stuff isn't easy to reuse. Etc.
To solve the problem we serially create templating libraries. Scads of them. Why so many? My guess is that they're never quite satisfactory. By the time they start getting powerful enough they've turned into monstrosities. Granted, some of them - such as SQLAlchemy and Genshi in the python world - are very beautiful and admirable monstrosities. Let's... um... avoid mention of PHP.
Because strings make an awkward interface between the worked-on language and the worked-with, we create a third language - templates - to avoid them. ** This also tends to be a little awkward.
Now let's look at a block of quoted Lisp code:
'(loop for i from 1 to 8 do (print i))
What do you see? As a new Lisp coder, I've caught myself looking at that as a string. It isn't. It is inactive Lisp code. You are looking at a bunch of lists and symbols. Try to evaluate it after turning one of the parentheses around. The language won't let you do it: syntax is enforced.
Using quasiquote, we can shoehorn our own values into this inactive Lisp code:
`(loop for i from 1 to ,whatever do (print i))
Note the nature of the shoehorning: one item has been replaced with another. We aren't formatting our value into a string. We're sliding it into a slot in the code. It's all neat and tidy.
In fact if you want to directly edit the text of the code, you are in for a hassle. For example if you are inserting a name <varname> into the code, and you also want to use <varname>-tmp in the same code you can't do it directly like you can with a template string: "%s-tmp = %s". You have to extract the name into a string, rewrite the string, then turn it into a symbol again and finally insert.
If you want to grasp the essence of Lisp, I think that you might gain more by ignoring defmacro and gensyms and all that window dressing for the moment. Spend some time exploring the potential of the quasiquote, including the ,# thing. It's pretty accessible. Defmacro itself only provides an easy way to execute the result of quasiquotes. ***
What you should notice is that the hermetic string/template barrier between the worked-on and the worked-with is all but eliminated in Lisp. As you use it, you'll find that your sense of two distinct layers - active and passive - tends to dissipate. Functions call macros which call macros or functions which have functions (or macros!) passed in with their arguments. It's kind of a big soup - a little shocking for the newcomer. That said, I don't find that the distinction between macros and functions is as seamless as some Lisp people say. Mostly it's ok, but every so often as I wander in the soup I find myself bumping up against the ghost of that old barrier - and it really creeps me out!
I'll get over it, I'm sure. No matter. The convenience pays for the scare.
Now that's Lisp working on Lisp. What about working on other languages? I'm not quite there yet, personally, but I think I see the light at the end of the tunnel. You know how Lisp people keep going on about S-expressions being the same thing as a parse tree? I think the idea is to parse the foreign language into S-expressions, work on them in the amazing comfort of the Lisp environment, then send them back to native code. In theory, every language out there could be turned into S-expressions, or even executable lisp code. You're not working in a first language combined with a third language to produce code in a second language. It is all - while you are working on it - Lisp, and you can generate it all with quasiquotes.
Have a look at this (borrowed from PCL):
(define-html-macro :mp3-browser-page ((&key title (header title)) &body body)
`(:html
(:head
(:title ,title)
(:link :rel "stylesheet" :type "text/css" :href "mp3-browser.css"))
(:body
(standard-header)
(when ,header (html (:h1 :class "title" ,header)))
,#body
(standard-footer))))
Looks like an S-expression version of HTML, doesn't it? I have a feeling that Lisp works just fine as its own templating library.
I've started to wonder about an S-expression version of python. Would it qualify as a Lisp? It certainly wouldn't be Common Lisp. Maybe it would be nicer - for python programmers at least. Hey, and what about P-expressions?
* Python now has something called AST, which I haven't explored. Also a person could use python lists to represent other languages. Relative to Lisp, I suspect that both are a bit of a hack.
** SQLAlchemy is kind of an exception. It's done a nice job of turning SQL directly into python. That said, it appears to have involved significant effort.
*** Take it from a newbie. I'm sure I'm glossing over something here. Also, I realize that quasiquote is not the only way to generate code for macros. It's certainly a nice one, though.
Data is code is an interesting paradigm that supports treating a data structure as a command. Treating data in this way allows you to process and manipulate the structure in various ways - e.g. traversal - by evaluating it. Moreover, the 'data is code' paradigm obviates the need in many cases to develop custom parsers for data structures; the language parser itself can be used to parse the structures.
The first step is forgetting everything you have learned with all the C and Pascal-like languages. Empty your mind. This is the hardest step.
Then, take a good introduction to programming that uses Lisp. Don't try to correlate what you see with anything that you know beforehand (when you catch yourself doing that, repeat step 1). I liked Structure and Interpretation of Computer Programs (uses Scheme), Practical Common Lisp, Paradigms of Artificial Intelligence Programming, Casting Spels in Lisp, among others. Be sure to write out the examples. Also try the exercises, but limit yourself to the constructs you have learned in that book. If you find yourself trying to find, for example, some function to set a variable or some statement that resembles a for loop, repeat step 1, then revisit the chapters before to find out how it is done in Lisp.
Read and understand the legendary page 13 of the Lisp 1.5 Programmer's Manual
According to Alan Kay, at least.
One of the reasons that some university computer science programs use Lisp for their intro courses is that it's generally true that a novice can learn functional, procedural, or object-oriented programming more or less equally well. However, it's much harder for someone who already thinks in procedural statements to begin thinking like a functional programmer than to do the inverse.
When I tried to pick up Lisp, I did it "with a C accent." set! amd begin were my friends and constant companions. It is surprisingly easy to write Lisp code without ever writing any functional code, which isn't the point.
You may notice that I'm not answering your question, which is true. I merely wanted to let you know that it's awfully hard to get your mind thinking in a functional style, and it'll be an exciting practice that will make you a stronger programmer in the long run.
Kampai!
P.S. Also, you'll finally understand that "my other car is a cdr" bumper sticker.
To truly grok lisp, you need to write it.
Learn to love car, cdr, and cons. Don't iterate when you can recurse. Start out writing some simple programs (factorial, list reversal, dictionary lookup), and work your way up to more complex ones (sorting sets of items, pattern matching).
On the code is data and data is code thing, I wouldn't worry about it at this point. You'll understand it eventually, and its not critical to learning lisp.
I would suggest checking out some of the newer variants of Lisp like Arc or Clojure. They clean up the syntax a little and are smaller and thus easier to understand than Common Lisp. Clojure would be my choice. It is written on the JVM and so you don't have the issues with various platform implementations and library support that exist with some Lisp implementations like SBCL.
Read On Lisp and Paradigms in Artificial Intelligence Programming. Both of these have excellent coverage of Lisp macros - which really make the code is data concept real.
Also, when writing Lisp, don't iterate when you can recurse or map (learn to love mapcar).
it's important to see that data is code AND code is data. This feeds the eval/apply loop. Recursion is also fun.
(This link is broken:
![Eval/Apply][1]
[1]: http://ely.ath.cx/~piranha/random_images/lolcode-eval_apply-2.jpg
)
I'd suggest that is a horrible introduction to the language. There are better places to start and better people/articles/books than the one you cited.
Are you a programmer? What language(s)?
To help you with your question more background might be helpful.
About the whole "code is data" thing:
Isn't that due to the "von Neumann architecture"? If code and data were located in physically separate memory locations, the bits in the data memory could not be executed whereas the bits in the program memory could not be interpreted as anything but instructions to the CPU.
Do I understand this correctly?
I think to learn anything you have to have a purpose for it, such as a simple project.
For Lisp, a good simple project is a symbolic differentiator, so for example
(diff 'x 'x) -> 1
(diff 'a 'x) -> 0
(diff `(+ ,xx ,yy) 'x) where xx and yy are subexpressions
-> `(+ ,(diff xx 'x),(diff yy 'x))
etc. etc.
and then you need a simplifier, such as
(simp `(+ ,x 0)) -> x
(simp `(* ,x 0)) -> 0
etc. etc.
so if you start with a math expression, you can eval it to get its value, and you can eval its derivative to get its derivative.
I hope this illustrates what can happen when program code manipulates program code.
As Marvin Minsky observed, computer math is always worried about accuracy and roundoff error, right? Well, this is either exactly right or completely wrong!
You can get LISP in many ways, the most common is by using Emacs or working next to somebody who has developed LISP already.
Sadly, once you get LISP, it's hard to get rid of it, antibiotics won't work.
BTW: I also recommend The Adventures of a Pythonista in Schemeland.
This may be helpful: http://www.defmacro.org/ramblings/fp.html (isn't about LISP but about functional programming as a paradigm)
The way I think about it is that the best part of "code is data" is the face that function are, well, functionally no different than another variable. The fact that you can write code that writes code is one of the single most powerful (and often overlooked) features of Lisp. Functions can accept other functions as parameters, and even return functions as a result.
This lets one code at a much higher level of abstraction than, say, Java. It makes many tasks elegant and concise, and therefore, makes the code easier to modify, maintain, and read, or at least in theory.
I would say that the only way to truly "get" Lisp is to spend a lot of time with it -- once you get the hang of it, you'll wish you had some of the features of Lisp in your other programming languages.