It's of course risky to discuss something that hasn't shipped, though extrapolating PNaCl's behavior from its architecture isn't that hard. (It's just NaCl except that LLVM is compiled to NaCl on load. And LLVM itself runs within the NaCl sandbox too.)
Off the top of my head, if PNaCl existed today, its advantages would be: access to SIMD units, true threads, 64-bit integer operations, 32-bit floats, debuggability with nacl-gdb, a growable heap, support for exceptions/setjmp, and no ridiculousness like: https://mail.mozilla.org/pipermail/es-discuss/2013-March/029...
As a browser implementer, you are fully justified in rejecting Pepper. :) Permanent APIs should go through some kind of standardization process. (Or they should HAVE, cough XHR)
I gave a presentation at GDC this morning* on Emscripten and asm.js and agree that asm.js has better survivability characteristics and is likely to succeed in the market. Nonetheless, the proposed design of PNaCl has several technical merits over asm.js as it exists today.
In the meantime, please add 64-bit integers, threads, and SIMD to asm.js :)
* It occupied the same time slot as the Mozilla presentation! arg!
Edit: forgot exceptions, setjmp, and a growable heap, none of which work in asm.js today. I know, I know, asm.js is early, but the point I'm making is that we're comparing two theoretical technologies at this point.
SIMD is coming to standard JS, and quickly in prototype form in SpiderMonkey. The idea that we need Dart (a whole new language and VM) or PNaCL + Pepper to get SIMD in JS is silly. Just add SIMD to JS, less work!
Some of what you rightly want is thus pure JS extension, not "asm.js". Other things (sbrk/mmap) are asm.js. Both can make progress, and cross-browser. Pepper cannot. That demerit trumps any SIMD-in-NaCl merit.
Right, but until they're in the language spec and implemented across browsers, you can't use 64-bit integer value objects in asm.js code unless you can polyfill or generate asm.js code that only runs in the latest Mozilla or whatever.
I AM using the same yardstick (neither asm.js nor PNaCl is usable today if you want the things I listed, but I predict PNaCl will exist in some form sooner), and I'm asking you to at least talk about the merits of PNaCl. Instead of constantly beating the JavaScript drum, you could have instead said "We have tried to work with Google to standardize PNaCl without Pepper and they are not open to discussion."
I know you have objections to Pepper, and I don't disagree, but personally I would be happy with a PNaCl minus Pepper too. Make WebIDL accessible from PNaCl and we can ignore all of the API issues.
I have a great deal of respect for you, Brendan, and Mozilla, and this is purely a communication tone issue, but as a game developer, it frequently sounds like Mozilla is more interested in the purity of the web than actually solving developer use cases. In fact, I had a couple discussions with game developers just yesterday about the fact that it feels like Google cares more.
HOWEVER, asm.js, parts of ES6/7, WebGL, etc. go a long way towards meeting the needs of game developers, so I know you actually do care about solving use cases. But you could at least be more respectful of Google's direct approach.
Personally, as someone who works with LLVM IR on a daily basis, the idea of standardizing LLVM IR as it exists today worries me a lot for several reasons:
* LLVM IR wasn't really designed to be portable. clang bakes lots of architecture-specific stuff into the IR that it generates. It can be hacked around the way PNaCl does by trying to define a "generic" architecture, but it's stretching the architecture beyond what it was intended to do.
* LLVM IR is very unstable and continually changes from version to version.
If we don't care about backwards compatibility, I would much rather see a bytecode standard done from scratch than PNaCl's approach of trying to use LLVM IR. I know the reason they did it is to use LLVM's code generation backends, but I feel like that could be done without using a large underspecified compiler IR as the transport format.
Sorry, it was you, not me, who shifted the topic from "merit" to "doesn't work cross-browser". I called that out because I'm pretty sure int64 and uint64, along with SIMD intrinsics, are coming to cross-browser JS -- while NaCl and PNaCl and in particular Pepper are not going cross-browser.
If you want to stick to merit-only arguments, please don't knock int64 as Mozilla-only or we'll circle back to the relative odds of JS standards evolving vs. infeasibility of standardizing Pepper + *NaCl.
On tone, I hope the above makes clear why I responded to your "not in other browsers" comment.
On purity, Mozilla has been faulted for being impure forever (not GPL, not XML/XHTML, not forsaking the average mass-market user for elite developers, etc. etc.). Our asm.js work is fully pragmatic and our talk at GDC (sorry it overlapped yours) had exactly the right tone. We are here to help game devs, without any purity tax.
On NaCl (not PNaCl) minus Pepper, I think we agree. We're looking at zerovm and similar uses of SFI in the context of the Servo project, for safer native code. That is where NaCl really shines, IMHO: safer native and OS/CPU-targeted code in native apps. It would be a shame if NaCl doesn't cater to this better use-case and instead tilts at the PNaCl cross-browser windmill.
How about evaluating both (asm.js and PNaCl)? Looks like we are stuck in the "monoculture" you said you dislike. And this IS the reason I never love web standards. I don't like monocultures too.
I feel you are biased to preserve JavaScript because you are the creator of JavaScript. IMO, asm.js might beat Flash but has a clear dead-end to beat native. We all are in a local optimum trap, I think.
Do we all buy two significantly different and expensive cars to "evaluate" both? Most people can't afford to do any such thing.
Nor do we all buy the same make and model of car. That's implementation monoculture.
Monoculture of implementation (WebKit, which I pointed out is fragmented in practice) is not the same as a single vocabulary for each kind of language at the primitive level of the web standards layer cake.
Yes, we are stuck with the web platform bottoming out at a certain set of languages. That's not monoculture, rather: termination without too much redundancy in something not so complex that multiple browsers can hope to implement interoperably. The platform has to stand on certain primitives.
Remember XHTML2 vs. HTML5 back in 2005? It was going to be one or the other, even if browsers such as Firefox parsed both (but XHTML2 slowly and without incremental layout).
Maybe a Microsoft or Google could have "done both" well (but IE didn't, except by treating text/xhtml as error-corrected HTML!).
Without a single dominant browser vendor with enough cash and willpower to pull off more than the minimum set of semantically sufficient/complementary primitive languages, we were likely stuck with HTML/CSS/JS. SVG and MathML as HTML5 polyglot languages eventually were rationalized into the stack but still see too little use to be optimized well (and didn't Chrome turn off MathML recently?).
So the primitives have to be minimized against competing browsers' ability to implement them all interoperably. Browsers weren't going to do equally well at both the 2005-era XML languages (XHTML/SVG/XSLT) and the competing incremental plan (HTML5).
Arbitrary native code has been falling out as a part of the primitive layer for a while. Plugins such as Flash and Silverlight are in decline. You could say we had a plugin polyculture, but it was a security nightmare and also bad for users just in terms of updates and cross-browser/plugin UX consistency.
Pepper is an attempt to mediate plugins' access to OS and browser APIs, but too tied to chromium and too large for Google to standardize or for other vendors to swallow.
Consider also how Dart support (2nd VM, in particular GC barriers required for cross-VM-heap cycle collection) bounced off WebKit:
I'm mostly being descriptive so far, but let me prescribe:
Based on the hard multi-vendor specification and software engineering problems of doing well with a roughly minimized set of primitives that we have today, I do not believe it's "better" for anyone, especially for web developers, to over-extend the primitive set with things like NaCl and Pepper.
That applies even if one vendor could, e.g., add SIMD primitives quickly. That just fragments the picture for developers, and misdirects some energy that could be spent on adding SIMD to JS.
Same goes for Dart. It was so important to do as a "replacement" (see leaked Dash memo) and not wait for the standards bodies, but that started in 2010 (or earlier? I knew only for certain in spring 2010). It's 2013 now and Dart is not going cross browser. Things it wanted, e.g. bignums, could have been put in ES6 by now and implemented in leading JS engines.
I do not agree we are in a local optimum trap, not with Unreal 3 running within 2x of native unsafe code speed.
> Pepper is an attempt to mediate plugins' access to OS and browser APIs, but too tied to chromium and too large for Google to standardize or for other vendors to swallow.
You keep repeating this, but I don't see how you can be remotely serious about it.
Compared to JS, DOM, HTML, and CSS, Pepper is downright simple:
As an external developer, Pepper is something I could actually implement independently for a reasonable amount of money ($2M or less) in a reasonable amount of time (1 year or less), if I could rely on Google's existing (P)NaCL toolchain.
Would there be ambiguity? Yes. Google has already said they're willing to work with implementors to resolve any design or ambiguity issues.
On the other hand, I can't even imagine trying to write a full browser stack, including HTML(5), CSS, JS, et al, in one year for <= $2M? Not a chance.
Yet you call Pepper "too big" and "too complex". Meanwhile, the enormous weight of the mess that is the web stack creates a huge barrier to entry for anyone that might want to meaningfully contribute new ideas to the space.
We're already seeing interesting development occur around (the much more approachable) NaCL/Pepper:
> That just fragments the picture for developers, and misdirects some energy that could be spent on adding SIMD to JS.
This is rich coming from the CTO that's refused to collaborate with Google and instead is going off and "misdirecting" energy on alternative approaches.
> It's 2013 now and Dart is not going cross browser.
You know, you might have something to do with that? You're not exactly a powerless or disinterested third party in this dialog.
> As an external developer, Pepper is something I could actually implement independently for a reasonable amount of money ($2M or less) in a reasonable amount of time (1 year or less), if I could rely on Google's existing (P)NaCL toolchain.
But what's the point? You'd be stuck running some subset of "things that can run on the web", and the vast expanse of existing HTML content would be inaccessible. If you want to use NaCL/PNaCl for non-web applications nobody is stopping you. Writing a "browser" that doesn't implement HTML+JS is just nonsense.
> Would there be ambiguity? Yes. Google has already said they're willing to work with implementors to resolve any design or ambiguity issues.
We have no basis to know whether "attempt to specify all out the ambiguitity in the Pepper API, using the Chromium implementation as the standard" is harder than "implement the already-fairly-well-specified set of web technologies". We know web technologies can be implemented fairly interoperably, we have multiple interoperable implementations.
> You know, you might have something to do with that? You're not exactly a powerless or disinterested third party in this dialog.
And yet we're not the only game in town, either. None of the other browser manufacturers has expressed any interest, so it's not like we're alone on this. Why should we have to carry Google's banner? We've got enough stuff on our plate.
> But what's the point? You'd be stuck running some subset of "things that can run on the web", and the vast expanse of existing HTML content would be inaccessible. If you want to use NaCL/PNaCl for non-web applications nobody is stopping you. Writing a "browser" that doesn't implement HTML+JS is just nonsense.
Unless you consider the possibility that:
- Web apps versus web documents are two very different things
- You see this as a way by which we could escape the mess and massive size of the current web stack by standardizing on a much smaller base.
- Simplifying the core stack could open the door to new competition and innovation in the field.
Personally, I'd love to see WebKit and/or Gecko and all the browser chrome implemented on top of NaCL and/or asm.js, with only the most minimal runtime, such that you're not sitting in a privileged position of running native code while the rest of us are stuck in JS/HTML/CSS hell.
> We know web technologies can be implemented fairly interoperably, we have multiple interoperable implementations.
We also know that it's ridiculously expensive and difficult, in no small part due to the resistance to enforcing things like formal validation of HTML in the browser itself, creating a problem that's as much art as it is science.
> And yet we're not the only game in town, either. None of the other browser manufacturers has expressed any interest, so it's not like we're alone on this. Why should we have to carry Google's banner? We've got enough stuff on our plate.
Apple and Microsoft have no interest in undercutting their own vertically integrated platforms.
Only you and Google have an strong interest in furthering the web, and yet, you refuse to work with Google despite them expending considerable time and money on furthering the web, while sharing their work freely.
We at Mozilla don't have $2M and a year to spare, but you're wrong: we've assessed doing Pepper with high fidelity (not porting chromium and duplicating all the DOM and other browser API implementations). It's more like $10M and multiple elapsed years to get to Chrome parity, assuming Chrome doesn't keep evolving and put us on a treadmill.
But then, I'm just the guy managing engineering and worrying about budget at Mozilla. Maybe you have greater skills. Where do you work?
Anyway, Pepper is big, with over a thousand methods among all the interfaces that are "specified" only by C++ implementation code we cannot port. We have a DOM implementation already, for example. So you cannot escape the fact that Pepper is "and also", not "instead of" -- there's no savings, it is purely added-cost, and significant cost.
I'm almost the only guy who will say this on HN, but as far as I can tell, Microsoft and Apple are on the same page. Maciej Stachowiak of Apple has agreed on HN, for what it's worth:
Also, Epic is just the first big game publisher we are working with. Epic folks have confirmed again just this week that they won't do NaCl, there's no benefit to something CWS-only. But they are super-excited about cross-browser web-wide reach with asm.js and browser-hosted C++-sourced games. That's the PNaCl output format and runtime for Google to collaborate on.
> It's more like $10M and multiple elapsed years to get to Chrome parity, assuming Chrome doesn't keep evolving and put us on a treadmill.
So you don't believe Google's offers to work with external implementors?
> But then, I'm just the guy managing engineering and worrying about budget at Mozilla. Maybe you have greater skills. Where do you work?
I run a systems programming consulting shop. We focus on low-level OS/driver and VM projects, with application development making up the remainder of our work.
In theory, reliable estimates are my livelihood, and I have a very vested interest in the future of the platforms that we will have to develop for. I'm very concerned about the market requiring us to develop apps for a Mozilla platform.
> Microsoft and Apple are on the same page
What reason do they have to roll out something that helps Google undercut their vertical platform/app markets?
If NaCL is 'too good', then it threatens their business. Take that for what it means in terms of asm.js optimization adoption.
> Epic folks have confirmed again just this week that they won't do NaCl, there's no benefit to something CWS-only.
CWS-only and Chrome-only are again, oddly, something you have a hand in. Google pushed out beta PNaCL tools, and CWS-only has primarily been (rightfully) pending PNaCL.
> So you don't believe Google's offers to work with external implementors?
I wasn't born yesterday.
Microsoft is patching WebKit to support Pointer Events. Strange days, but even this "help" does not mean Pointer Events should win (or lose -- it's neutral in standards terms, at most helpful to show implementation feasibility and quality, if possible).
Why do you distrust Apple and Microsoft and ascribe bad motives to them based on their commercial interests, but fault Mozilla for not jumping on Google's Pepper treadmill like a good little-brother caricature? Google has commercial interests too, and they have spent >$1B a year on Chrome -- including advertising and bundling that directly targets Firefox users. (It's amazing we are still alive, and possibly even growing desktop share.)
You are surely right that some amount of competitive pressure will be required to get asm.js optimized in all engines. Ditto for WebGL in IE, and enabled in Safari. But cooperation between Chrome and Firefox is already happening on both asm.js and WebGL, so between these "coopetition" pincers, I bet we'll prevail.
Anyway, let's assume everyone has equally good motives, since Google is not any less commercial than Microsoft or Apple these days. Pepper is still too costly for others to swallow even with offers of "help", and therefore losing -- it is not even in the race.
asm.js and evolved Web APIs (which developers need anyway) are winning.
Evil-me on my Throne of Skulls didn't ordain this outcome. It is unfolding before our eyes across browsers, developers, and very pragmatic third parties including big game/game-engine publishers.
(But I am laughing a Dr. Evil laugh, on my Skull island. :-P)
> Why do you distrust Apple and Microsoft and ascribe bad motives to them based on their commercial interests, but fault Mozilla for not jumping on Google's Pepper treadmill like a good little-brother caricature?
I don't distrust them, or ascribe bad motives. I ascribe motives that align with their economic interests -- it doesn't matter if they're bad or good.
> Google has commercial interests too, and they have spent >$1B a year on Chrome -- including advertising and bundling that directly targets Firefox users. (It's amazing we are still alive, and possibly even growing desktop share.)
How much does Google spend on Firefox's search deal? What do you see as Google's commercial motive with Chrome?
My reading (in the context that I have some family that works on Chrome) is that their primary interest was ensuring that they could help push the web forward, and not be beholden to external interests in doing so. They have an interest in open platforms insofar as it gives them leverage over the much more strongly established platform players (eg, Microsoft, Apple).
> asm.js and evolved Web APIs (which developers need anyway) are winning.
The interesting thing is that if asm.js has sufficient adoption, the 'web' part of the equation may very well not matter at all.
If Google maintains their efforts on (P)NaCL and can provide better performance, end implementors may target both asm.js and NaCL, with shims for platform-specific functionality (not unlike most game development today). Once you're doing that, you can also target other native platforms with the same code and tooling.
So I don't think asm.js is all bad. Ironically, it could very well be the escape route from JS/HTML/CSS/DOM that we've all been praying for ever since the idea of the web app came into being. Not because it's the best technical solution, but since when has technical correctness outweighed market forces?
> Evil-me on my Throne of Skulls didn't ordain this outcome.
Of course not. But "evil you" certainly has a hand in it. There are only 4 real players in the browser market, and you're one of them.
In fact, you're the one that decided to ignore what Google has been doing and invest in asm.js to begin with, so divesting any claim in the role of asm.js's possible ascendency is a bit of a stretch.
My use of "bad" and "good" about your relative judgments of Microsoft, Apple, and Google could be "purple" and "shiny" for all it matters.
The point is you seem to treat Google as more benign and non-commercial. That's not credible based on many pieces of evidence, including their public company performance scrutiny by Wall Street. I admire Google for some things they do, and believe they do not over-optimize their share price, but they cannot ignore it, either. Also, they are a house divided, with many conflicting agendas not directly related to business goals.
> How much does Google spend on Firefox's search deal?
That's not something I can comment on, per our contract stipulated by Google, but the rumors are online and if you believe them, they show a commercial partnership, nothing more.
Numerate folks have estimated the value and cost of various search deals, see e.g. Jeremy Wagstaff of Reuters. I won't comment, except to say that Mozilla was underpaid for a long time, something we chose early on in order to avoid being greedy and triggering a bad reaction from search partners.
> What do you see as Google's commercial motive with Chrome?
Lots of motives, some mixed. It's complex, and the "make the web better" motive is still there and all to the good. Some shift away from standardization toward "works in Chrome/CWS" -- and not due to anything I did -- is evident lately, and disturbing. At the limit, it's Microsoft-y.
Again, if Google as a whole were to standardize early and often, just to take one example as we've done with the missing device and sensor APIs for mobile via Firefox OS (with Samsung patching WebKit for Tizen to match), we'd have a better web, faster. Some of the delays there can be blamed on Android, but not all.
> The interesting thing is that if asm.js has sufficient adoption, the 'web' part of the equation may very well not matter at all.
No, for Emscripten/ASM.js, you still need a C or C++ runtime, not just libc/stdio/stdlib stuff but various graphics, audio, and other APIs.
> In fact, you're the one that decided to ignore what Google has been doing and invest in asm.js to begin with, so divesting any claim in the role of asm.js's possible ascendency is a bit of a stretch.
Here you show your bias. I didn't "ignore" what Google has been doing, I estimated it as too costly to risk.
Now you tell me why the shoe isn't on the other foot. Why did Google "ignore" what we've been doing to advance JS for the last two years, and move all its Aarhus talent onto Dart, at some cost to V8? And at the cost of Epic, a "sale" we won that Google could have, had it only invested a bit more in JS.
You really do have a pro-Google, anti-Mozilla animus -- good/bad or purple/shiny, I don't care.
> The point is you seem to treat Google as more benign and non-commercial. That's not credible based on many pieces of evidence, including their public company performance scrutiny by Wall Street. I admire Google for some things they do, and believe they do not over-optimize their share price, but they cannot ignore it, either. Also, they are a house divided, with many conflicting agendas not directly related to business goals.
Absolutely not; I consider Google to be an organization whose commercial interests are ultimately far more aligned with my own than Apple's or Microsoft's. It has nothing to do with being benign or non-commercial -- my interests are in no small part commercial.
> No, for Emscripten/ASM.js, you still need a C or C++ runtime, not just libc/stdio/stdlib stuff but various graphics, audio, and other APIs.
Sure, but we (the systems/games development community) are well equipped to create common runtime libraries, especially on a platform that provides an anemic set of platform standards and components, in which case custom libraries won't suffer from QT or Swing's uncanny valley problem.
> Now you tell me why the shoe isn't on the other foot. Why did Google "ignore" what we've been doing to advance JS for the last two years, and move all its Aarhus talent onto Dart, at some cost to V8?
It seems to me that Google made a rational decision that JavaScript is not a technologically sound foundation upon which to build, and made the first moves to find alternatives.
> You really do have a pro-Google, anti-Mozilla animus -- good/bad or purple/shiny, I don't care.
It's actually an anti-JavaScript animus. The idea that JavaScript ought to be the foundation of the next generation of modern computing platforms runs counter to everything I know about language and runtime design.
It seems to be an entirely market-driven solution, not a technologically-driven one -- but those market circumstances are something you have far more influence over than I do.
You just wrote that I "ignored" Google's NaCl and Pepper work. That is not true, and yet you didn't hold Google to account anywhere on this thread (or others) for "ignoring" opportunities such as asm.js which were latent in Emscripten (in spite of their commercial support for Mandreel-compiled CWS games).
There's more to what you wrote than just anti-JS animus. But let's not get stuck on this dispute; let's be unstuck like JS :-P.
> It seems to be an entirely market-driven solution, not a technologically-driven one -- but those market circumstances are something you have far more influence over than I do.
Look deeper. The market is responding to cost structures ultimately based on how hard it is to do PNaCl + Pepper vs. Emscripten + asm.js + (Web APIs that devs need anyway).
This is not about "best tech" or "marketing" or "market power". (Firefox doesn't have Google's marketing budget, and we definitely do not have decisive market power, as Netscape and then IE did -- but neither does Chrome.)
This is about "shortest path" or "worse is better" -- who gets adoption first with something "good enough" wins, and can keep winning if the good-enough thing keeps evolving and doesn't get stuck.
JS is "good enough" and manifestly capable of evolving cross-browser via the Ecma TC39 standards process and some coopetition to be even better.
Anyway, at the asm.js level, apart from syntax aesthetics, why do you care how an int32 cast or type annotation is encoded? The semantics are sufficient, as proven by UE3 in Firefox. They only get richer as we do SIMD, task ||ism, etc.
> This is about "shortest path" or "worse is better ...
I think we're in agreement there.
> JS is "good enough" and manifestly capable of evolving cross-browser via the Ecma TC39 standards process and some coopetition to be even better.
Perhaps in disagreement there, especially when you're competing against Apple's user-first focus, and especially when I'm on the receiving end of being forced to use a JS-based stack as a developer. I do mean the full stack, too -- instruction level profiling, debugging, failure reporting, consistent performance profiles, the whole nine yards.
> Anyway, at the asm.js level, apart from syntax, why do you care how an int32 cast or type annotation is encoded? The semantics are sufficient, as proven by UE3 in Firefox. They only get richer as we do SIMD, task ||ism, etc.
Well, as an implementor of systems software, I can't help but be wary that asm.js brings along with it the entire JS language and (in practice) web stack, too. It means that the likelihood that 'web' applications will ever be free of the enormous web stack is lowered, though alternative implementations (eg, NaCL) reduce some of that risk.
I share some concern about Apple, but less so in this "post Apple" era (Google, Samsung, up-and-comers from China).
On asm.js, the subset is verified at parse time and also as needed at link time, so there's no taint of "full JS" -- yet. We do expect the heaps to cross-connect, and want that. It won't make the parts of JS you dislike live much longer, from what I can tell.
To be perfectly candid, my goal is to evolve standardized JS, and therefore its implemented and widely distributed VMs, to be a good multi-language target language (and runtime).
This is eminently doable and closer than many people think. If it succeeds, then JS-the-handcoded-source-language can live or die on its (as evolved by then) language merits, and I won't cry if it dies.
But I wouldn't bet on its fan base dropping it even if they can choose other languages.
Off the top of my head, if PNaCl existed today, its advantages would be: access to SIMD units, true threads, 64-bit integer operations, 32-bit floats, debuggability with nacl-gdb, a growable heap, support for exceptions/setjmp, and no ridiculousness like: https://mail.mozilla.org/pipermail/es-discuss/2013-March/029...
As a browser implementer, you are fully justified in rejecting Pepper. :) Permanent APIs should go through some kind of standardization process. (Or they should HAVE, cough XHR)
I gave a presentation at GDC this morning* on Emscripten and asm.js and agree that asm.js has better survivability characteristics and is likely to succeed in the market. Nonetheless, the proposed design of PNaCl has several technical merits over asm.js as it exists today.
In the meantime, please add 64-bit integers, threads, and SIMD to asm.js :)
* It occupied the same time slot as the Mozilla presentation! arg!
Edit: forgot exceptions, setjmp, and a growable heap, none of which work in asm.js today. I know, I know, asm.js is early, but the point I'm making is that we're comparing two theoretical technologies at this point.