atto...@etang.com (attonie) writes: > as i know, lisp is a interpretion language, thus the speed is slow, > so is there any program can compile lisp program to machine code ?
Where do you 'know' that from? Every Common Lisp implementation nowadays has a compiler, most compile directly to machine code, some to C or bytecode. There are even implementations which don't have an 'interpreter' at all (but nevertheless allow for interactive development as if there were an interpreter).
* atto...@etang.com (attonie) | as i know, lisp is a interpretion language, thus the speed is slow, so | is there any program can compile lisp program to machine code?
You must be thinking of Scheme, a toy Lisp. Common Lisp, a real Lisp, has always had compilers. Matter of fact, even in 1959, Lisp had a compiler. It is just that the toy Lisps that are used in education and other forms of confusion are implemented by people who want to prove some academic point -- and these all use Scheme. There is significant merit to learning about computers from a higher-level point of view than, say, machine instructions and stuff, and Scheme offers a way to look at Lambda calculus. More of this is visible with interpreted code than with compiled machine language, so for purpose of study, Scheme is usually an interpreted Lisp.
Lisp has, however, always provided an interactive programming environment and many toolmakers in the history of computing have made interactive environments using interpreted languages, because it is _so_ much easier than writing compilers that work with interactive environments, but all this means is that people who are not very smart never gave you the power of an interactive environment combined with the speed of compiled machine code. The people who make Common Lisp environments are much smarter than those who stay with interpreted environments, so they provide you with this particular combination as a matter of course.
Now, over the years, a lot of people who have been overexposed to Scheme and to lazy or stupid people and their toy implementations of Scheme have come to Common Lisp with the severely misguided notions that Lisp is slow, Lisp is interpreted, Lisp is a memory hog, etc, all because they have gathered all their experience about Lisp from incompetent or lazy people. What you have actually seen is that even incompetent and lazy people can make some useful Lisp environments that can be really nifty compared to the compile-only environments of other toy languages. This should be an important clue. It is, sadly, often missed completely.
It is like those LEGO robot kits for modern kids (of any age :), with which you can do a lot of fascinating stuff with building blocks and a computer -- but it is of course no match for industrial robots, and anyone who went into an automated machine shop and thought he knew how robots were made, would be laughed at and kicked out. The nature of the "virtual reality" on the Net provides no clues to what most people need visual information to get, namely that their assumptions are all wrong. When you start an obvious falsehood with "as I know", it it very easy to think you are one of those guys whose assumptions are set in stone and for whom reality had better fit nicely with them, or else. So it is very easy to laugh at you and kick you out because of this. However, it is up to you whether you accept that what you have known up to now is wrong (or applies only to a very small fragment of reality, not to be confused with larger fragments or the whole thing) and are prepared to listen to people who know a lot more about this than you do, or whether you will insist that you had a right to believe what you did, which a lot of people will not grant you, and be laughed and tried to be kicked out.
If you really want the toy Lisp, go ask in comp.lang.lisp.toy, probably known at your newsserver as comp.lang.scheme. Some Scheme freaks also offer compiled versions, whatever the point of that waste would be. As you might infer, Scheme is the annoying little kid brother who always tags along and ruins your day, and Scheme freaks have an annyoing habit of trying to make it on their own, only to blame real Lisp when it fails. I do not think you _really_ want the toy Lisp, do you? :)
Visit www.lisp.org and look for links to commercial vendors who offer their excellent environments in trial versions or to free, but much smaller environments that require a slightly more educated user to enjoy from the start. Note that these are not "apps", but environments.
/// -- In a fight against something, the fight has value, victory has none. In a fight for something, the fight is a loss, victory merely relief.
* attonie wrote: > as i know, lisp is a interpretion language, thus the speed is slow, so > is there any program can compile lisp program to machine code ?
As far as I know there is one current common lisp (clisp) which doesn't compile to native machine code but to a virtual `bytecode' machine like Java. There is also a rather old system, originally Kyoto Common Lisp, now GNU Common Lisp (gcl) which compiles to C (and thence to native code). All the other implementations I know of compile to native code. So your problem is rather the inverse of what you think: to find a Common Lisp system which does not compile to machine code. If you count bytecode as machine code, then I don't think there are any such systems.
In article <3223529447153...@naggum.net>, Erik Naggum wrote: >* atto...@etang.com (attonie) >| as i know, lisp is a interpretion language, thus the speed is slow, so >| is there any program can compile lisp program to machine code?
> You must be thinking of Scheme, a toy Lisp. Common Lisp, a real Lisp,
attonie wrote: > as i know, lisp is a interpretion language, thus the speed is slow, so > is there any program can compile lisp program to machine code ?
As this is probably the most unartfully constructed troll I have seen in a while, I will give an answer. ..
Since you KNOW that Lisp is an interpretation language, why would you even suspect that a program to compile Lisp to machine code would (or even could) exist? In fact, how do you KNOW that Lisp exists?
Seek within what you KNOW for what you DON'T KNOW, but BELIEVE. In there, the answer shall lie.
My what big teeth you have. Give the guy a break, his English isn't so good, and I got the feeling that he meant, 'as far as I know'. Lisp being interpreted only is a pretty common misconception, even propogated by some textbooks talking about only early Lisp. If he was a going to spread FUD, why do it in the Lisp ng? Isn't it more likely he was seeking information? Perhaps this is a good time to give useful information about CL, and end his misconception instead of giving this ng a hostile atmosphere.
> > as i know, lisp is a interpretion language, thus the speed is slow, so > > is there any program can compile lisp program to machine code ?
> As this is probably the most unartfully constructed troll I have seen in a > while, I will give an answer. ..
> Since you KNOW that Lisp is an interpretation language, why would you even > suspect that a program to compile Lisp to machine code would (or even > could) exist? In fact, how do you KNOW that Lisp exists?
> Seek within what you KNOW for what you DON'T KNOW, but BELIEVE. In there, > the answer shall lie.
http://www.bagley.org/~doug/shootout/craps.shtml Look at what those benchmarks show, CMUCL scored just behind C, C++, and a couple ML implementations (but ahead of other ML's). It should be noted that these were algorithms designed to work with all languages, so Lisp didn't get an adequate chance to show its muscle.
CL is /not/ a purely interpreted language, it is one of the few languages with true dynamic compilation. That is much more impressive than it is in other languages because of macros, which none of the ML's have. Plus, type declarations are optional, in ML they are most certainly not.
> > attonie wrote:
> > > as i know, lisp is a interpretion language, thus the speed is slow, so > > > is there any program can compile lisp program to machine code ?
> Since you KNOW that Lisp is an interpretation language, why would you even > suspect that a program to compile Lisp to machine code would (or even > could) exist? In fact, how do you KNOW that Lisp exists?
This isn't fair. `As I know' was fairly obviously `As far as I know' and I'm fairly sure that this form of it is what some non-native speakers produce - I know a Portuguese/Brazilian person who I think produces this unless he's thinking hard.
Also, given that a lot of courses which `teach' (or mention) Lisp are horribly inaccurate in just this way I think this person should be given the benefit of the doubt - they may just have been mislead and genuinely looking for enlightenment.
Tim Bradshaw <t...@cley.com> writes: > * attonie wrote: > > as i know, lisp is a interpretion language, thus the speed is slow, so > > is there any program can compile lisp program to machine code ?
> As far as I know there is one current common lisp (clisp) which > doesn't compile to native machine code but to a virtual `bytecode' > machine like Java. There is also a rather old system, originally > Kyoto Common Lisp, now GNU Common Lisp (gcl) which compiles to C (and > thence to native code).
Just to clarify. The latest and more supported Kyoto Common Lisp derivative is ECL.
Cheers
-- Marco Antoniotti ======================================================== NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488 719 Broadway 12th Floor fax +1 - 212 - 995 4122 New York, NY 10003, USA http://bioinformatics.cat.nyu.edu "Hello New York! We'll do what we can!" Bill Murray in `Ghostbusters'.
> * atto...@etang.com (attonie) > | as i know, lisp is a interpretion language, thus the speed is slow, so > | is there any program can compile lisp program to machine code?
> You must be thinking of Scheme, a toy Lisp. Common Lisp, a real Lisp, > has always had compilers. Matter of fact, even in 1959, Lisp had a > compiler. It is just that the toy Lisps that are used in education and > other forms of confusion are implemented by people who want to prove some > academic point -- and these all use Scheme. There is significant merit > to learning about computers from a higher-level point of view than, say, > machine instructions and stuff, and Scheme offers a way to look at Lambda > calculus. More of this is visible with interpreted code than with > compiled machine language, so for purpose of study, Scheme is usually an > interpreted Lisp.
Just to be clear, there are compilers for Scheme, including:
MzScheme Kawa (Scheme to JVM bytecode) Skij (Scheme to JVM bytecode) Stalin Chicken Bigloo Scheme-to-C Gambit Scheme etc.
In fact, I think there are more implementations of Scheme compilers than Common Lisp compilers. Which is not to say Scheme is preferable to CL (or vice versa); it's just that Scheme (too) has compilers.
Alexander Schofield wrote: > My what big teeth you have. > Give the guy a break, his English isn't so good, and I got the feeling > that he meant, 'as far as I know'. Lisp being interpreted only is a > pretty common misconception, even propogated by some textbooks talking > about only early Lisp. If he was a going to spread FUD, why do it in > the Lisp ng? Isn't it more likely he was seeking information? Perhaps > this is a good time to give useful information about CL, and end his > misconception instead of giving this ng a hostile atmosphere.
Well, I've been around `this group for a good many years. Most people coming in here with a conception of Lisp as an "interpreted-only" language have either not done ANY research (try a google search on Lisp compiler and you'll find plenty of hits) or *is* trolling. As such, I felt somewhat justified in responding as I did. As for the perception that this NG is a "hostile" place, well, not all of the world is pretty, but I daresay I've not even been close to being the thorniest thistle about here. In any case, if my coments were cause for offense, rather than insight, I do apologize for the form in which I presented it. The semantics, however, stand.
