I am confused about JavaScript engines right now. I know that V8 was a big deal because it compiled JavaScript to native code.
Then I started reading about Mozilla SpiderMonkey, which from what I understand is written in C and can compile JavaScript. So how is this different from V8 and if this is true, why does Firefox not do this?
Finally, does Rhino literally compile the JavaScript to Java byte code so you would get all the speed advantages of Java? If not, why do people not run V8 when writing scripts on their desktops?
There are various approaches to JavaScript execution, even when doing JIT. V8 and Nitro (formerly known as SquirrelFish Extreme) choose to do a whole-method JIT, meaning that they compile all JavaScript code down to native instructions when they encounter script, and then simply execute that as if it was compiled C code. SpiderMonkey uses a "tracing" JIT instead, which first compiles the script to bytecode and interprets it, but monitors the execution, looking for "hot spots" such as loops. When it detects one, it then compiles just that hot path to machine code and executes that in the future.
Both approaches have upsides and downsides. Whole-method JIT ensures that all JavaScript that is executed will be compiled and run as machine code and not interpreted, which in general should be faster. However, depending on the implementation it may mean that the engine spends time compiling code that will never be executed, or may only be executed once, and is not performance critical. In addition, this compiled code must be stored in memory, so this can lead to higher memory usage.
The tracing JIT as implemented in SpiderMonkey can produce extremely specialized code compared to a whole-method JIT, since it has already executed the code and can speculate on the types of variables (such as treating the index variable in a for loop as a native integer), where a whole-method JIT would have to treat the variable as an object because JavaScript is untyped and the type could change (SpiderMonkey will simply "fall off" trace if the assumption fails, and return to interpreting bytecode). However, SpiderMonkey's tracing JIT currently does not work efficiently on code with many branches, as the traces are optimized for single paths of execution. In addition, there's some overhead involved in monitoring execution before deciding to compile a trace, and then switching execution to that trace. Also, if the tracer makes an assumption that is later violated (such as a variable changing type), the cost of falling off trace and switching back to interpreting is likely to be higher than with a whole-method JIT.
V8 is the fastest, because it compiles all JS to machine code.
SpiderMonkey (what FF uses) is fast too, but compiles to an intermediate byte-code, not machine code. That's the major difference with V8. EDIT- Newer Firefox releases come with a newer variant of SpideMonkey; TraceMonkey. TraceMonkey does JIT compilation of critical parts, and maybe other smart optimizations.
Rhino compiles Javascript into Java classes, thus allowing you to basically write "Java" applications in Javascript. Rhino is also used as a way to interpret JS in the backend and manipulate it, and have complete code understanding, such as reflection. This is used for example by the YUI Compressor.
The reason why Rhino is used instead of V8 all over the place is probably because V8 is relatively new, so a lot of projects have already been using Rhino/Spidermonkey as their JS engine, for example Yahoo widgets. (I assume that's what you're referring to with "scripts on their desktops")
edit-
This link might also give some insight of why SpiderMonkey is so widely adopted.
Which Javascript engine would you embed in your application?
If you want to see how the various in-browser Javascript engines stack up, install Safari 4 (yes it runs on Windows now too!), Chrome V8, Firefox 3.5, and IE 8 (if you are on windows) and run the benchmark:
http://www2.webkit.org/perf/sunspider-0.9/sunspider.html
I believe as Pointy said above, the new Firefox 3.5 uses TraceMonkey that also compiles to intermedit code on the fly using some form of JIT. So it should compare to V8 somewhat favorably. At least it won't be 10x slower than V8 like Firefox 3 SpiderMonkey (without JIT) was.
For me... safari 4.0.3 was 2.5x faster than Tracemonky in Firefox 3.5.3 on Win XP. IE8 was much much slower. I don't have Chrome installed at the moment.
Don't know about Rhino compiling to java bytecode. If it's still interpreting the dynamic features of Javascript such as being able to add attributes to object instances at runtime (example obj.someNewAttribute="someValue" which is allowed in Javascript)... I'm not so sure that it's entirely "compiled" to bytecode, and you might not get any better performance other than you don't have to compile from Javascript source code text every time your Javascript runs. Remember that Javascript allows very dynamic syntax such as eval("x=10;y=20;z=x*y"); which means you can build up strings of code that get compiled at runtime. That's why I'd think Rhino would be mixed-mode interpreted/compiled even if you did compile to JVM bytecode.
The JVM is still an interpreter, albeit a very good one with JIT support. So I like to think of Rhino-on-JVM as 2 interpreter layers (interpreter-on-interpreter) or interpreter^2. Whereas most of your other Javascript engines are written in C, and as such should perform more like interpreter^1. Each interpreter layer can add 5-10x performance degradation as compared to C or C++ (ref Perl or Python or Ruby for example), but with JIT the performance hit can be much lower on the order of 2-4x. And the JVM has one of the most robust & mature JIT engines ever.
So your mileage will definitely vary and you probably would benefit from doing some serious benchmarks if you want a real answer for your intended application on your own hardware & OS.
Rhino can't be too awfully slow, since I know lots of folks are using it. I think it's main appeal is not its speed, but the fact that is easy-to-code/light-weight/embeddable/interpreter that has hooks into Java libraries, which makes it perfect for scripting/configuration/extensibility of your software project. Some text editors like UltraEdit are even embedding Javascript as an alternative macro scripting engine. Every programmer seems to be able to stumble through javascript pretty easily, so it's easy to pick up as well.
One advantage to Rhino is that it runs pretty much anywhere the JVM runs. In my experience, trying to get stand-alone TraceMonkey or SpiderMonkey to build & run from command line can be a bit painful on systems like Windows. And embedding in your own application can be even more time consuming. But the payback to having an embeddable language would be worth it for a big project, as compared to having to "roll your own" mini scripting solution if that's what you're looking to do.
Scripting with Rhino is really easy if you have Java and the rhino jar, you just write your javascript and run it from command line. I use it all the time for simple tasks.
To respond the question, why native code Vs Byte code...
The native code is faster and for google a strategic choice because they have plan to JS one of them at least is ChromeOS.
A good video about this question is posted on Channel9 with an interview with Lars Bak the man behind V8 can be found here
2022 Update
V8 as also SpiderMonkey do support MultiPass Compilation as also SinglePass Compilation.
Java Rhino got deprecated and replaced by a new concept called Truffle Framework it is part of the GraalVM Stack which is in fact a JVMCI added to the JVM where CI means Compiler Interface. That allows the Truffle Language Implementer Framework to Implement any Language and translate that to Java Byte Code in MultiPass and Single Pass. the ECMAScript implementation is called GraalJS and it comes in 2 Flavors
GraalJS the 1:1 Rhino Replacement including compatability mode for incremental adoption and graal-node a NodeJS Fork where v8 is replaced with the GraalJS Engine.
as sayed Truffle is a Polyglot Language Implementer Framework so that works with any GraalVM Supported Language.
hope that helps and makes some sense since the jar versions do most best work when the whole stack has aligned versions you should alaways install the complet stack:
https://github.com/graalvm/graalvm-ce-builds/releases
download and install that then use the so called gu command to add graaljs
follow the instructions here: https://github.com/oracle/graaljs
Yes GraalVM can in general Compile Single Binary Executeable Files of all Polyglot Languages this is called AOT Compilation it has faster boot times but less throughput it can also do JIT and it can do a mix of both.
Would it not make sense to support a set of languages (Java, Python, Ruby, etc.) by way of a standardized virtual machine hosted in the browser rather than requiring the use of a specialized language -- really, a specialized paradigm -- for client scripting only?
To clarify the suggestion, a web page would contain byte code instead of any higher-level language like JavaScript.
I understand the pragmatic reality that JavaScript is simply what we have to work with now due to evolutionary reasons, but I'm thinking more about the long term. With regard to backward compatibility, there's no reason that inline JavaScript could not be simultaneously supported for a period of time and of course JavaScript could be one of the languages supported by the browser virtual machine.
Well, yes. Certainly if we had a time machine, going back and ensuring a lot of the Javascript features were designed differently would be a major pastime (that, and ensuring the people who designed IE's CSS engine never went into IT). But it's not going to happen, and we're stuck with it now.
I suspect, in time, it will become the "Machine language" for the web, with other better designed languages and APIs compile down to it (and cater for different runtime engine foibles).
I don't think, however, any of these "better designed languages" will be Java, Python or Ruby. Javascript is, despite the ability to be used elsewhere, a Web application scripting language. Given that use case, we can do better than any of those languages.
I think JavaScript is a good language, but I would love to have a choice when developing client-side web applications. For legacy reasons we're stuck with JavaScript, but there are projects and ideas looking for changing that scenario:
Google Native Client: technology for running native code in the browser.
Emscripten: LLVM bytecode compiler to javascript. Allows LLVM languages to run in the browser.
Idea: .NET CLI in the browser, by the creator of Mono: http://tirania.org/blog/archive/2010/May-03.html
I think we will have JavaScript for a long time, but that will change sooner or later. There are so many developers willing to use other languages in the browser.
Answering the question - No, it would not make sense.
Currently the closest things we have to a multi-language VM are the JVM and the CLR. These aren't exactly lightweight beasts, and it would not make sense to try and embed something of this size and complexity in a browser.
Let's examine the idea that you could write a new, multilanguage VM that would be better than the existing solution.
You're behind on stability.
You're behind on complexity (way, way, behind because you're trying to generalize over multiple languages)
You're behind on adoption
So, no, it doesn't make sense.
Remember, in order to support these languages you're going to have to strip down their APIs something fierce, chopping out any parts that don't make sense in the context of a browser script. There are a huge number of design decisions to be made here, and a huge opportunity for error.
In terms of functionality, we're probably only really working with the DOM anyway, so this is really an issue of syntax and language idom, at which point it does make sense to ask, "Is this really worth it?"
Bearing in mind, the only thing we're talking about is client side scripting, because server side scripting is already available in whatever language you like. It's a relatively small programming arena and so the benefit of bringing multiple languages in is questionable.
What languages would it make sense to bring in? (Warning, subjective material follows)
Bringing in a language like C doesn't make sense because it's made for working with metal, and in a browser there isn't much metal really available.
Bringing in a language like Java doesn't make sense because the best thing about it is the APIs anyway.
Bringing in a language like Ruby or Lisp doesn't make sense because JavaScript is a powerful dynamic language very close to Scheme.
Finally, what browser maker really wants to support DOM integration for multiple languages? Each implementation will have its own specific bugs. We've already walked through fire dealing with differences between MS Javascript and Mozilla Javascript and now we want to multiply that pain five or six-fold?
It doesn't make sense.
On Windows, you can register other languages with the Scripting Host and have them available to IE. For example VBScript is supported out of the box (though it has never gained much popularity as it is for most purposes even worse than JavaScript).
The Python win32 extensions allowed one to add Python to IE like this quite easily, but it wasn't really a good idea as Python is quite difficult to sandbox: many language features expose enough implementation hooks to allow a supposedly-restricted application to break out.
It is a problem in general that the more complexity you add to a net-facing application like the browser, the greater likelihood of security problems. A bunch of new languages would certainly fit that description, and these are new languages that are also still developing fast.
JavaScript is an ugly language, but through careful use of a selective subset of features, and support from suitable object libraries, it can generally be made fairly tolerable. It seems incremental, practical additions to JavaScript are the only way web scripting is likely to move on.
I would definitely welcome a standard language independent VM in browsers (I would prefer to code in a statically typed language).
(Technically) It's quite doable gradually: first one major browser supports it and server has the possibility to either send bytecode if current request is from compatible browser or translate the code to JavaScript and send plain-text JavaScript.
There already exist some experimental languages that compile to JavaScript, but having a defined VM would (maybe) allow for better performance.
I admit that the "standard" part would be quite tricky, though. Also there would be conflicts between language features (eg. static vs. dynamic typing) concerning the library (assuming the new thing would use same library). Therefore I don't think it's gonna happen (soon).
If you feel like you are getting your hands dirty, then you have either been brainwashed, or are still feeling the after affects of the "DHTML years". JavaScript is very powerful, and is suited well for its purpose, which is to script interactivity client side. This is why JavaScript 2.0 got such a bad rap. I mean, why packages, interfaces, classes, and the like, when those are clearly aspects of server-side languages. JavaScript is just fine as a prototype-based language, without being full-blown object oriented.
If there is a lack of seamlessness to your applications because the server-side and client-side are not communicating well, then you might want to reconsider how you architect your applications. I have worked with extremely robust Web sites and Web applications, and I have never once said, "Hmm, I really wish JavaScript could do (xyz)." If it could do that, then it wouldn't be JavaScript -- it would be ActionScript or AIR or Silverlight. I don't need that, and neither do most developers. Those are nice technologies, but they try to solve a problem with a technology, not a... well, a solution.
I don't think that a standard web VM is that inconceivable. There are a number of ways you could introduce a new web VM standard gracefully and with full legacy support, as long as you ensure that any VM bytecode format you use can be quickly decompiled into javascript, and that the resulting output will be reasonably efficient (I would even go so far as to guess that a smart decompiler would probably generate better javascript than any javascript a human could produce themselves).
With this property, any web VM format could be easily decompiled either on the server (fast), on the client (slow, but possible in cases where you have limited control of the server), or could be pre-generated and loaded dynamically by either the client or the server (fastest) for browsers that don’t natively support the new standard.
Those browsers that DO natively support the new standard would benefit from increased speed of the runtime for web vm based apps. On top of that, if browsers base their legacy javascript engines on the web vm standard (i.e. parsing javascript into the web vm standard and then running it), then they don’t have to manage two runtimes, but that’s up to the browser vendor.
While Javascript is the only well-supported scripting language you can control the page directly from, Flash has some very nice features for bigger programs. Lately it has a JIT and can also generate bytecode on the fly (check out runtime expression evaluation for an example where they use flash to compile user-input math expressions all the way to native binary). The Haxe language gives you static typing with inference and with the bytecode generation abilities you could implement almost any runtime system of your choice.
Quick update on this old question.
Everyone who affirmed that a "web page would contain byte code instead of any higher-level language like JavaScript" "won't happen".
June 2015 the W3C announced WebAssembly that is
a new portable, size- and load-time-efficient format suitable for
compilation to the web.
This is still experimental, but there is already some prototypal implementation in Firefox nightly and Chrome Canary and there is already some demonstration working.
Currently, WebAssembly is mostly designed to be produced from C/C++, however
as WebAssembly evolves it will support more languages than C/C++, and we hope that other compilers will support it as well.
I let you have a closer look at the official page of the project, it is truly exciting!
this question resurfaces regularly. my stance on this is:
A) wont happen and B) is already here.
pardon, what? let me explain:
ad A
a VM is not just some sort of universal magical device. most VMs are optimized for a certain language and certain language features. take the JRE/Java (or LLVM): optimized for static typing, and there are definitely problems and downsides when implementing dynamic typing or other things java didn't support in the first place.
so, the "general multipurpose VM" that supports lots of language features (tail call optimization, static & dynamic typing, foo bar boo, ...) would be colossal, hard to implement and probably harder to optimize to get good performance out of it. but i'm no language designer or vm guru, maybe i'm wrong: it's actually pretty easy, only nobody had the idea yet? hrm, hrm.
ad B
already here: there may not be a bytecode compiler/vm, but you don't actually need one. afaik javascript is turing complete, so it should be possible to either:
create a translator from language X to javascript (e.g. coffeescript)
create a interpreter in javascript that interprets language X (e.g. brainfuck). yes, performance would be abysmal, but hey, can't have everything.
ad C
what? there wasn't a point C in the first place!? because there isn't ... yet. google NACL. if anyone can do it, it's google. as soon google gets it working, your problems are solved. only, uh, it may never work, i don't know. the last time i read about it there were some unsolved security problems of the really tricky kind.
apart from that:
javascript's been there since ~1995 = 15 years. still, browser implementations differ today (although at least it's not insufferable anymore). so, if you start something new yet, you might have a version working cross browser around 2035. at least a working subset. that only differs subtly. and needs compatibility libs and layers. no point in not trying to improve things though.
also, what about readable source code? i know a lot of companies would prefer not to serve their code as "kind-of" open source. personally, i'm pretty happy i'm able to read the source if i suspect something fishy or want to learn from it. hooray for source code!
Indeed. Silverlight is effectively just that - a client side .Net based VM.
There are some errors in your reasoning.
A standard virtual machine in a standard browser will never be standard. We have 4 browsers, and IE has conflicting interests with regard to 'standard'. The three others are evolving fast but adoption rate of new technologies is slow. What about browsers on phones, small devices, ...
The integration of JS in the different browsers and its past history leads you to under-estimating the power of JS. You pledge a standard, but disapprove JS because standard didn't work out in the early years.
As told by others, JS is not the same as AIR/.NET/... and the like. JS in its current incarnation perfectly fits its goals.
In the long term, Perl and Ruby could well replace javascript. Yet the adoption of those languages is slow and it is known that they will never take over JS.
How do you define best? Best for the browser, or best for the developer? (Plus ECMAScript is different than Javascript, but that is a technicality.)
I find that JavaScript can be powerful and elegant at the same time. Unfortunately most developers I have met treat it like a necessary evil instead of a real programming language.
Some of the features I enjoy are:
treating functions as first class citizens
being able to add and remove functions to any object at any time (not useful much but mind blowing when it is)
it is a dynamic language.
It's fun to deal with and it is established. Enjoy it while it is around because while it may not be the "best" for client scripting it is certainly pleasant.
I do agree it is frustrating when making dynamic pages because of browser incompatibilities, but that can be mitigated by UI libraries. That should not be held against JavaScript itself anymore than Swing should be held against Java.
JavaScript is the browser's standard virtual machine. For instance, OCaml and Haskell now both have compilers that can output JavaScript. The limitation is not JavaScript the language, the limitation is the browser objects accessible via JavaScript, and the access control model used to ensure you can safely run JavaScript without compromising your machine. The current access controls are so poor, that JavaScript is only allowed very limited access to browser objects for safety reasons. The Harmony project is looking to fix that.
It's a cool idea. Why not take it a step further?
Write the HTML parser and layout engine (all the complicated bits in the browser, really) in the same VM language
Publish the engine to the web
Serve the page with a declaration of which layout engine to use, and its URL
Then we can add features to browsers without having to push new browsers out to every client - the relevant new bits would be loaded dynamically from the web. We could also publish new versions of HTML without all the ridiculous complexity of maintaining backwards compatibility with everything that's ever worked in a browser - compatibility is the responsibility of the page author. We also get to experiment with markup languages other than HTML. And, of course, we can write fancy JIT compilers into the engines, so that you can script your webpages in any language you want.
I would welcome any language besides javascript as possible scripting language.
What would be cool is to use other languages then Javascript. Java would probably not be a great fit between the tag but languages like Haskell, Clojure, Scala, Ruby, Groovy would be beneficial.
I came a cross Rubyscript somewhile ago ...
http://almaer.com/blog/running-ruby-in-the-browser-via-script-typetextruby and http://code.google.com/p/ruby-in-browser/
Still experimental and in progress, but looks promising.
For .Net I just found: http://www.silverlight.net/learn/dynamic-languages/ Just found the site out, but looks interesting too. Works even from my Apple Mac.
Don't know how good the above work in providing an alternative for Javascript, but it looks pretty cool at first glance. Potentially, this would allow one to use any Java or .Net framework natively from the browser - within the browser's sandbox.
As for safety, if the language runs inside the JVM (or .Net engine for that matter), the VM will take care of security so we don't have to worry about that - at least not more then we should for anything that runs inside the browser.
Probably, but to do so we'd need to get the major browsers to support them. IE support would be the hardest to get. JavaScript is used because it is the only thing you can count on being available.
The vast majority of the devs I've spoken to about ECMAScript et. al. end up admitting that the problem isn't the scripting language, it's the ridiculous HTML DOM that it exposes. Conflating the DOM and the scripting language is a common source of pain and frustration regarding ECMAScript. Also, don't forget, IIS can use JScript for server-side scripting, and things like Rhino allow you to build free-standing apps in ECMAScript. Try working in one of these environments with ECMAScript for a while, and see if your opinion changes.
This kind of despair has been going around for some time. I'd suggest you edit this to include, or repost with, specific issues. You may be pleasantly surprised by some of the relief you get.
A old site, but still a great place to start: Douglas Crockford's site.
Well, we have already VBScript, don't we? Wait, only IE supports it!
Same for your nice idea of VM. What if I script my page using Lua, and your browser doesn't have the parser to convert it to bytecode? Of course, we could imagine a script tag accepting a file of bytecode, that even would be quite efficient.
But experience shows it is hard to bring something new to the Web: it would take years to adopt a radical new change like this. How many browsers support SVG or CSS3?
Beside, I don't see what you find "dirty" in JS. It can be ugly if coded by amateurs, propagating bad practice copied elsewhere, but masters shown it can be an elegant language too. A bit like Perl: often looks like an obfuscated language, but can be made perfectly readable.
Check this out http://www.visitmix.com/Labs/Gestalt/ - lets you use python or ruby, as long as the user has silverlight installed.
This is a very good question.
It's not the problem only in JS, as it is in the lack of good free IDEs for developing larger programs in JS. I know only one that is free: Eclipse. The other good one is Microsoft's Visual Studio, but not free.
Why would it be free? If web browser vendors want to replace desktop apps with online apps (and they want) then they have to give us, the programmers, good dev tools. You can't make 50,000 lines of JavaScript using a simple text editor, JSLint and built-in Google Chrome debugger. Unless you're a macohist.
When Borland made an IDE for Turbo Pascal 4.0 in 1987, it was a revolution in programming. 24 years have passed since. Shamefully, in the year 2011 many programmers still don't use code completion, syntax checking and proper debuggers. Probably because there are so few good IDEs.
It's in the interest of web browser vendors to make proper (FREE) tools for programmers if they want us to build applications with which they can fight Windows, Linux, MacOS, iOS, Symbian, etc.
Realistically, Javascript is the only language that any browsers will use for a long time, so while it would be very nice to use other languages, I can't see it happening.
This "standardised VM" you talk of would be very large and would need to be adopted by all major browsers, and most sites would just continue using Javascript anyway since it's more suited to websites than many other browsers.
You would have to sandbox each programming language in this VM and reduce the amount of access each language has to the system, requiring a lot of changes in the languages and removal or reimplementation of many features. Whereas Javascript already has this in mind, and has done a for a long time.
Maybe you're looking for Google's Native Client.
In a sense, having a more expressive language like Javascript in the browser instead of something more general like Java bytecode has meant a more open web.
I think this is not so easy issue. We can say that we're stuck with JS, but is it really so bad with jQuery, Prototype, scriptaculous, MooTools, and all fantastic libraries?
Remember, JS is lightweight, even more so with V8, TraceMonkey, SquirrelFish - new Javascript engines used in modern browsers.
It is also proved - yeah, we know it has problems, but we have lots of these sorted out, like early security problems. Imaging allowing your browser to run Ruby code, or anything else. Security sandbox would have to be done for scratch. And you know what? Python folks already failed two times at it.
I think Javascript is going to be revised and improved over time, just like HTML and CSS is. The process may be long, but not everything is possible in this world.
I don't think you "understand the pragmatic issue that JavaScript is simply what we have to work with now". Actually it is very powerful language. You had your Java applet in browser for years, and where is it now?
Anyhow, you don't need to "get dirty" to work on client. For example, try GWT.
... you mean...
Java and Java applet
Flash and Adobe AIR
etc..
In general, any RIA framework can fill your needs; but for every one there's a price to pay for using it ( ej. runtime avalible on browser or/and propietary or/and less options than pure desktop )
http://en.wikipedia.org/wiki/List_of_rich_internet_application_frameworks
For developing Web with any non-web languaje, you've GWT: develop Java, compile to Javascript
Because they all have VMs with bytecode interpreters already, and the bytecode is all different too. {Chakra(IE), Firefox (SpiderMonkey), Safari (SquirrelFish), Opera(Carakan).
Sorry , I think Chrome (V8) compiles down to IA32 machine code.
well, considering all browsers already use a VM, I don't think it will be that difficult to make a VM language for the web.
I think it would greatly help for a few reasons:
1. since the server compiles the code, the amount of data sent is smaller and the client doesn't waist time on compiling the code.
2. since the server can compile the code in preparation and store it, unlike the client which tries to waist as little time quickly compiling the JS, it can make better code optimizations.
3. compiling a language to byte code is way easier then transpiling to JS.
as a final note (as someone already said in another comment), HTML and CSS compile down to a simpler language, not sure if it counts as byte code, but you could also send compiled html and css from the server to the client which would reduce parse and fetch times
IMO, JavaScript, the language, is not the problem. JavaScript is actually quite an expressive and powerful language. I think it gets a bad rep because it's not got classical OO features, but for me the more I go with the prototypal groove, the more I like it.
The problem as I see it is the flaky and inconsistent implementations across the many browsers we are forced to support on the web. JavaScript libraries like jQuery go a long way towards mitigating that dirty feeling.