Brendan Eich on Ecma TC39, JavaScript.next, and Google Dash

I'd be curious to see this backed up. Certainly with a long enough timeline we are not at a maximum of "works only in x", and -- ignoring those beleaguered Opera users still subjected to crappy UA sniffing -- I can't think of the last time I came across a site that blocked me for whatever my choice of browser (I'm a Chrome and Firefox user as well).

If, on the other hand, you're talking about fancy new CSS3 or WebGL (or whatever) demos leaving out some browser due to neglecting to add a particular prefixed property (or whatever), I don't really see that as a problem. Prototype then standardize, right? Yes, not feature detecting and prefixing properly is poor web development practice and yes, they are cutting out part of their audience, but using an experimental feature inherently means that it may break at some point, and almost certainly isn't supported in all browsers.

When 3d CSS transforms land in Firefox soon, there are going to be a whole lot of demos out there that won't work. I would say that's an education and library problem, not a fragmentation one.

Google is doing deals and marketing web apps and add-ons as "Chrome". This is showing up all over the place, most recently in my experience on the Starbucks free wireless service agreement page.

Major content publishers Mozilla has talked to are surprised that we can host their Chrome-targeted web apps and HTML5-based extensions. They think such things are exclusive to Chrome. If they use Dart and need native performance, then that'll be true. There is a slippery slope, especially when more and more of the differentiation is not on a standards track (Dart, NaCl, WebSQL).

Nevertheless most of the current differentiation is bogus and any modern browser will do. There's certainly some "WebKit required" fragmentation. We saw this with Amazon's kindle web app, which is specifically built to WebSQL and not IndexedDB. But I do not believe any big player is pushing "WebKit required" as an agenda or marketing program. In contrast, Google is clearly pushing "Chrome" as a requirement or works-best-in brand.

The Chrome experiments, most recently

http://www.allisnotlo.st/index_en.html

are obvious, maybe too obvious. Defenders can always play up the "experiment" hedge. That helps but only so much, especially when many of these experiments could work in other browsers with trivial changes.

I'm told that people have run chrome.angrybirds.com in Firefox. So it should be html5.angrybirds.com but -- no surprise from a marketing angle -- the domain name starts with C not H.

First, I'm a huge fan. :)

... and I was with you until the WebSQL vs IndexedDB thing.

IndexedDB is a horrible, horrible thing. It should be put out of its misery and replaced with WebSQL in Firefox.

Mozilla is singlehandedly holding back offline web apps with the insistance on a slow and hard to program -- yet buzzword friendly -- technology that no one actually wants to use. There's no reason what so ever to stick with a dumbed down K/V store when you're not trying to replicate data across data centers.

So, IMHO, you've bought this bit of "WebKit required" onto yourselves. WebSQL is the only production ready offline web app technology available today.

Take it from someone who offline enabled one of the biggest web apps on the web: I'd rather leave out offline entirely than code IndexedDB.

Considering that Apple will never support it, it's pretty much dead. And I feel that Mozilla is really on the losing end of this particular battle.

TL:DR; Please support WebSQL in Firefox. :)

Without further nuance, saying "standards must keep up, deal with it" is a bit too consequentialist for me, a Mozilla founder, or for anyone working sincerely in the open web standards bodies. Korean and Chinese banks still require ActiveX PKI plugins or else deny service. The end of better web tech does not justify any means, especially not single-vendor locked-in means.

It seems to me Google is trying to have it both ways, which creates not just the appearance of a conflict, but an actual conflict: does their top talent work with standards bodies to make interoperable specs (based on prototypes and open source and so on), or do they go for wow-effect and sweet-talk if not strong-arm tactics?

Google knows they can't standardize something like NaCl. My belief is that control flow integrity enforcement will sooner come to OSes and OS-targeted toolchains than it will to browsers via NaCl and Pepper. So, safer (but still OS-specific) binary plugins, in the next five years. But mapping Pepper, an unimpressively large and messy API tied to Chrome as well as WebKit, onto other browsers? Other vendors won't get on that treadmill, and I don't believe Google would try to "do it for them". Gears was too painful.

On Dart: Google as a single entity does not know what could be done in Ecma TC39, since the Dart/V8 principals never participated, and V8 needed a fresh-thinking second team finally to get going on Harmony prototyping. But let's agree that one or two designers work better sprinting alone, not burdened by a committee (see my blog post for how we avoid design by committee).

Then Dart could come to Ecma as a proposed spec with a single open source implementation, perhaps later this year or early next. This was not part of the leaked game-plan, however. Clearly Google did not, as of that document's date, intend to standardize. And who believes they would give up "design control"?

Would TC39 entertain Dart as a second language? I doubt it. A whole second spec to write and get through ISO, cross-language-heap cycles to collect, mixed and possibly conflicting runtime semantics, specifically more numeric types in Dart than current JS to coerce and/or fail to convert in the native/managed bridges and API stubs, the list goes on.

Would TC39 take inspiration from Dart and fold ideas and design elements into the Harmony agenda? Absolutely we would, but we need to see Dart first, and this last week does not make for an auspicious launch. Still, no hard feelings if we do get a clean pitch from Dart principals to TC39.

However this plays out, Google has a lot of power. I won't quote Spider-Man's Uncle Ben, but for a company that moralizes about evil-as-in-don't-be, the standards they hold themselves to have to be high. I argue that they can't successfully and faithfully both work for interoperable and better web standards among multiple browsers, and play deep proprietary lock-in games. Pick one or the other.

Innovating in the open and proposing early and often to standards bodies are fine. Delayed-open-washing and increasing piles of proprietary (open-source doesn't matter) single-vendor code, which implements features that are not ever proposed for standardization, are not.

Even if any means are justified toward the end of improving web programmer productivity over what JS affords today, Dart represents a specific, clear and negative judgment on the Harmony work in Ecma, a judgment that I believe will be shown to be a mistake. It can't possibly help us make faster progress in TC39 -- it's at best a distraction and at worst a break in trust among the members.

With tooling (Brightly) and massive developer and evangelism resources backing Dart, will we really have the kind of open-web-first, fair-play contest that people who thought Google was (as Paul Graham put it) "aligned with the grain of the web" have come to want and expect from Google?

It looks like we won't. GOOG is acting more like MSFT of old (also like AOL, wanting sticky eyeballs and time-on-site instead of being the best search engine). The game theories of public companies, the innovator's dilemma, the Facebook challenge, and the browser-vendor Prisoner's Dilemma, all predict this ironic development. It is not a surprise.

But companies are made of people, and I have hopes that Googlers on the right side of this conflict will speak up.