Sponsored Link •
After a short introduction about the relevance of macros as tools to design abstractions, including full programming languages, I show some common patterns of Scheme macrology: recursive macros, accumulators, and the usage of literals to incorporate helpers in macros.
Macros are the reason why I first became interested in Scheme, five or six years ago. At the time - as at any time - there was a bunch of people trolling in comp.lang.python, arguing for the addition of macros to the language. Of course most Pythonistas opposed the proposal.
At the time I had no idea of the advantages/disadvantages of macros and I felt quite ignorant and powerless to argue. I never liked to feel ignorant, so I decided to learn macros, especially Scheme macros, because they are the state of the art for what concerns the topic.
Nowadays I have some arguments to back up the position against macros. I have two main objections, one technical (less important) and one political (more important).
The technical reason is that I do not believe in macros for languages without S-expressions. There are plenty of examples of macro systems without S-expressions - for instance Dylan or PLOT in the Lisp world and Logix and MetaPython in the Python world, but none of them ever convinced me. Scheme macros are much better because of the homoiconicity of the language ("homoiconicity" is just a big word for the code is data concept). [Notice that technically Scheme macros work on syntax objects and not directly on S-expressions like traditional Lisp macros, but this is a subtle point I will discuss when talking about hygiene; I can skip it for the moment being.]
I have already stated in episode 12 my political objection, i.e. my belief that macros have a high cost in terms of complication of the language (look how complicated the R6RS module system is!). Moreover, code based on macros tends to be too clever, difficult to debug, and sometime idiosyncratic; I do not want to maintain code such kind of code in a typical enterprise context, with programmers of any kind of competence. Sometimes I wish that even Python was a simpler language! There is a difference between simpler and dumber, of course. I am not implying that every enterprise should adopt only enterprise-oriented languages; as a matter of fact various cutting edge enterprises are taking advantage of non-conventional and/or research-oriented languages, but I see them as exceptions to the general rule.
My opinion is based on the fact that on my daily work (I use Python exclusively there) I have never felt the need for macros. For instance, I had occasion to write both small declarative languages and small command-oriented languages, but they were so simple that I had no need for Scheme macros. Actually, judging from my past experience, I think extremely unlikely that I will ever need something as sophisticated as Scheme macros in my daily work. The one thing that I miss in Python which Scheme has is pattern matching, not macros.
Having said that, I do not think that macros are worthless, and actually I think they are extremely useful and important in another domain, i.e. in the domain of design and research about programming languages. Scheme is certainly not the only language where you can experiment with language design, it is just the best language for this kind of tasks, at least in my humble opinion.
For instance, a few months ago I have described an experiment I did with the Python meta object protocol, in order to change how the object system work, and replace multiple inheritance with traits. Even if in Python it is possible to customize the object system, I do not thing the approach is optimal, because changing the semantics without changing the syntax does not feel right. In Scheme I could have implemented the same with a custom syntax and in a somewhat less magical way. I am interested with this kind of experiments, even if I will never use them in production code, and I use Scheme in preference for such purposes.
The major interest of Scheme macros for me lies in the fact that they enable every programmer to write her own programming language. I think this is a valuable thing. Anybody who has got opinions about language design, or about how an object system should should work, or questions like "what would a language look like if it had feature X?", can solve his doubts by implementing the feature with macros.
Notice that I recognize that perhaps not everybody should design its own programming language, and that certainly not everybody should distribute its own personal language. Nevertheless, I think everybody can have opinions about language design. Experimenting with macrology can help to put to test such opinions and to learn something.
The easiest approach is to start from a Domain Specific Language (DSL), which does not need to be a fully grown programming language. For instance, in the Python world everybody is implementing his own templating language to generate web pages. In my opinion, this a good thing per se, the problem is that everybody is distributing his own language so that there is a bit of anarchy.
Even for what concerns fully grown programming languages we see nowadays an explosion of new languages, especially for the Java and the .NET platforms, since it is relatively easy to implement a new language there. However, it still takes a substantial amount of work.
On the other hand, writing a custom language embedded in Scheme by means of macros is much easier. I see Scheme as an excellent platform for implementing languages and experimenting with new ideas.
There is a quote of Ian Bicking about Web frameworks which struck me:
Sometimes Python is accused of having too many web frameworks. And it's true, there are a lot. That said, I think writing a framework is a useful exercise. It doesn’t let you skip over too much without understanding it. It removes the magic. So even if you go on to use another existing framework (which I'd probably advise you do), you' ll be able to understand it better if you've written something like it on your own.
You can the replace the words "web framework" with "programming language" and the quote still makes sense. You should read my Adventures in this spirit: the ambition of the series is to give to the readers the technical competence to write small Scheme-embedded languages by means of macros. Even if you are not going to design your own language, macros will help you to understand how languages work.
Personally I am interested in the technical competence, I do not want to write a new language. There are already lots of languages out there, and writing a real language is a lot of grunt work, because it means writing debugging tools, good error messages, wondering about portability, interacting with an user community, et cetera et cetera.
The goal of learning macros well enough to implement a programming language is an ambitious one; it is not something I can attain in one episode of the Adventures, nor in six. However, one episode is enough to explain at least one useful technique which is commonly used in Scheme macrology and which is good to know in order to reach our final goal, in time.
The technique I will discuss in this episode is writing recursive macros with accumulators. In Scheme it is common to introduce an auxiliary variable to store a value which is passed in a loop - we discussed it in episode 6 when talking about tail call optimization: the same trick can be used in macros, at expand-time instead that at run-time.
In order to give an example I will define a macro cond minus (cond-) which works like cond, but with less parenthesis. Here is an example:
(cond- cond-1? return-1 cond-2? return-2 ... else return-default)
should expand to:
(cond (cond-1? return-1) (cond-2? return-2) ... (else return-default))
Here is a solution, which makes use of an accumulator and of an auxiliary macro cond-aux:
(def-syntax cond-aux (syntax-match () (sub (cond-aux (acc ...)) #'(cond acc ...)) (sub (cond-aux (acc ...) x1) #'(syntax-violation 'cond- "Mismatched pairs" '(acc ... x1) 'x1)) (sub (cond-aux (acc ...) x1 x2 x3 ...) #'(cond-aux (acc ... (x1 x2)) x3 ...)) )) (def-syntax (cond- x1 x2 ...) (cond-aux () x1 x2 ...))
The code above should be clear. The auxiliary macro cond-aux is recursive: it works by collecting the arguments x1, x2, ..., xn in the accumulator (acc ...). If the number of arguments is even, at some point we end up having collected all the arguments in the accumulator, which is then expanded into a standard conditional; if the number of arguments is even, at some point we end up having collected all the arguments except one, and a "Mismatched pairs" exception is raised. The user-visible macro cond- just calls cond-aux by setting the initial value of the accumulator to (). The entire expansion and error checking is made at compile time. Here is an example of usage:
> (let ((n 1)) (cond- (= n 1) ; missing a clause (= n 2) 'two (= n 3) 'three else 'unknown)) Unhandled exception: Condition components: 1. &who: cond- 2. &message: "Mismatched pairs" 3. &syntax: form: (((= n 1) (= n 2)) ('two (= n 3)) ('three else) 'unknown) subform: 'unknown
I have nothing against auxiliary macros, however sometimes you may want to keep all the code in a single macro. This is useful if you are debugging a macro since an auxiliary macro is usually not exported. The trick is to introduce a literal to defined the helper macro inside the main macro. Here is how it would work in this example:
(define-syntax cond- (syntax-match (aux) (sub (cond- aux (acc ...)) (cond acc ...)) (sub (cond- aux (acc ...) x1) (syntax-violation 'cond- "Mismatched pairs" '(acc ... x1) 'x1)) (sub (cond- aux (acc ...) x1 x2 x3 ...) (cond- aux (acc ... (x1 x2)) x3 ...)) (sub (cond- x1 x2 ...) (cond- aux () x1 x2 ...))))
If you do not want to use a literal identifier, you can use a literal string instead:
(define-syntax cond- (syntax-match () (sub (cond- "aux" (acc ...)) (cond acc ...)) (sub (cond- "aux" (acc ...) x) (syntax-violation 'cond- "Mismatched pairs" '(acc ... x) 'x)) (sub (cond- "aux" (acc ...) x1 x2 x3 ...) (cond- "aux" (acc ... (x1 x2)) x3 ...)) (sub (cond- x1 x2 ...) (cond- "aux" () x1 x2 ...))))
These tricks are quite common in Scheme macros: we may even call them design patterns. In my opinion the best reference detailing these techniques and others is the Syntax-Rules Primer for the Merely Eccentric, by Joe Marshall. The title is a play on the essay An Advanced Syntax-Rules Primer for the Mildly Insane by Al Petrofsky.
Marshall's essay is quite nontrivial, and it is intended for expert Scheme programmers. On the other hand, it is child play compared to Petrofsky's essay, which is intended for foolish Scheme wizards ;)
Have an opinion? Be the first to post a comment about this weblog entry.
|Michele Simionato started his career as a Theoretical Physicist, working in Italy, France and the U.S. He turned to programming in 2003; since then he has been working professionally as a Python developer and now he lives in Milan, Italy. Michele is well known in the Python community for his posts in the newsgroup(s), his articles and his Open Source libraries and recipes. His interests include object oriented programming, functional programming, and in general programming metodologies that enable us to manage the complexity of modern software developement.|