Hi everyone, I was searching the web these days trying to find out more about these two macro systems and understand their differences, and why one is preferable over the other (or not). I'd like to share with you some ideas, and hopefully get some opinions back as well. Coming from the lisp side, and without much knowledge on scheme macros, I'm also trying to decide for myself whether scheme macros can have a (practical) advantage.
One good point (found here http://lambdagrok.wikidot.com/forum/t-194636) I think explains it pretty well: lisp macros merely transform a list into another, while scheme macros use a pattern matching "sub- language" to "apply transformations" on the input syntax, and produce new syntax. I think this puts the finger on the spot. It seems to me that with macros, scheme breaks out of the homoiconicity of lisp, and opens up a new array of possibilities: define new languages with <b>different</b> syntax (while lisp allows new languages but with the same syntax). I'm saying this by looking at Racket and the direction it has chosen to go (have a look at Typed Scheme and Datalog). This looks like the world turned upside down: the pattern matching macro system is the essence of the system, and scheme is just another language defined with it. The list is not that important anymore, since it's not so essential for the macro system. Now, I have read some opinions which say that most who choose the scheme macros, make it for the pattern matching abilities and not for the hygienic part. This seems like a reasonable thing to do, since I don't hear lispers complain about unhygienicity. If there are people out there who had some practical experience with scheme macros, I hope they read this post and share some thoughts. I have a feeling that there's an additional gain with scheme macros: debugging information given by the system when something goes wrong. But it's only a feeling, I know too little of this subject to be sure. Macros are hard to debug. The stacktrace displayed by clojure does not contain information related to macro code. It's probably hard to do. My feeling relates to the fact that in scheme macro processing is not arbitrary, but rather strictly defined and under control. So I'm thinking that this gives scheme more control over what's going on in a macro and also enables scheme to give me more information in a stacktrace. I'd love to hear other opinions on this. Another point I find peculiar is the small attention that scheme macros got during all these years. I wonder why it's like that. It could be scheme's low score for practical stuff, but then again I don't know of another language that borrows these kind of macros. Does anybody know? And lastly, a question to Rich Hickey, should he read this post: what is the reasoning behind the choice of lisp macros over scheme's? Cheers, Razvan -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
