If I had to rank-order the various humiliations I've seen, this comes near the top.
Microsoft: "here's our best stab at what a great browser looks like - and it took us 14 years to get here".
Google: "we've been working on a codebase for 2 years. We decided to release it as a patch to your browser. Our brief foray into this area has improved your best effort by 10x".
True, but the guy who did V8 is the same guy who did Beta, Self, and HotSpot. That's like 20 years of experience in writing high-performance VMs for dynamic languages right there.
"Vision" isn't an easily defined quantity. What MS lacks is incentive to make a browser with fast JavaScript performance, as they make most of their money off desktop apps, many of which web-based apps are looking to replace.
So actually, the opposite is true: MS has a big economic incentive to slow down JS performance in browsers, because it will make them more money selling desktop softare. One way to do this is to release new versions of IE which have slow JS interpreters, and use marketing to convince enterprise customers that the new browsers are super-duper top-of-the-line shiny things, with features like security and easy maintainability.
That's a great business point which doesn't generally occur to developers who believe the world is a simpler place where good and bad software is created, but intentions are always sincere.
The internet is a killing blow for MSFT's business model, because it devalues all their proprietary assets. And when MSFT relents a bit to the internet and tries to play along, it does this with the same business model that sells software with closed standards (Silverlight)or is protected by patent (.net).
The browser represents a piece of software that enforces neither of these business model components that MSFT absolutely depends on, and it represents the door to computer interaction. This is really bad for MSFT as it exists today. MSFT will be a very different company 20 years from now, but I don't expect it will change without a serious fight given what's at stake. We shouldn't be surprised if MSFT engages in some practices which might be anathema to anybody just expecting a good software product from a company with substantial resources.
Microsoft has economic incentive to stay competitive on many fronts. I guarantee you that MSFT is working hard to make JS faster, to make their browser more standards compliant, and to substantially improve security. It makes no strategic or tactical sense for them to do otherwise -- to improve the browser is not to make the desktop less attractive.
I guarantee you that MSFT is working hard to make JS faster,
Then why haven't they made it as fast as a grad student volunteer did in 2 months for FireFox [1]? This kind of stuff has been pretty well researched in the past couple of decades :-P
It is incorrect to conclude that Microsoft's lack of achievement in X implies (1) its lack of ambition for improving X, or (2) [as ancestor suggested] its determination to sabotage X.
Is it really so hard to imagine what life on the inside looks like?
Microsoft is a huge institution. There are hundreds of engineers working on IE. Their work is segmented and bucketed. Some poor engineer has probably spent the last several years doing nothing but maintaining the EOT font format. There are reorganizations, new decisions to realign the next release to meet the goals of other teams, etc. Every decision -- at any level -- requires buy-in from multiple parties with overlapping responsibility. There are, in short, too many cooks in the kitchen.
I'm not saying it isn't ridiculous. But this is how it is inside Microsoft.
I know they have good people in there, but the fact is that they are pulling us back. They couldn't get off their asses to work on IE 7 until Firefox became a credible threat.
This is the most compelling evidence that Microsoft always thinks about preserving the status-quo instead of advancing the state of the art. It's their decision and their company, but I'm not going to accept apologies for them lightly.
They want to change their image, then they should play nicer with their competition and with us. I don't know how they could do that, but off the top of my head ... why not open-source Silverlight to make it a true standard? But they won't do that since they need control. And round and round we go.
> They couldn't get off their asses to work on IE 7 until Firefox became a credible threat.
This is very true and I've always felt it is a strong condemnation of what goes on inside Microsoft. I neglected to mention this sorry bit of history in my posts above mostly because it's a bit old news. Today, IE is a fully-staffed, heavily armed, etc.
I don't think I apologized for Microsoft. There is no excuse for having previously shut down IE, or for having such poor execution now that the IE team is back. There are, however, explanations... ones which hopefully help to dispel the myth that Microsoft is, today, intentionally holding the browser back. The truth is much more mundane.
First, it is entirely possible for Microsoft to continue to reap rewards from Windows while improving their browser at the same time. The two are not mutually exclusive. I reject the perspective that says the browser will win at the cost of desktop APIs. Everything has its place.
Second, while Windows may make shareholders (and your hypothetical "mid-level executives") fat today, they are under no illusions that it will make them fat tomorrow. Innovating for the future is an earnest and honest goal at Microsoft. This includes the browser, too. Who is to say where the next cash cow will be?
(Also: for some strange reason, I can't reply directly to you. Why would that be the case?)
Making the browser and web applications a viable alternative to their cash cows (Windows, Office, BackOffice) would seriously compromise the fat bonuses even mid-level executives get and would seriously jeopardize top management's position in front of their shareholders.
It simply does not happen this way. Not with a big public company that enjoy a monopoly with insanely fat margins.
And the guys who wrote the Squirrelfish VM for safari weren't all famous VM researchers. From what I read, they just picked all the low hanging fruit in classic VM design and got great performance - better than IE8 for sure. Their code is even open source. MS could just read through the source and copy the design.
Great things aren't going to happen by just having him on the payroll, Microsoft would have to make good use of his abilities by giving him resources, time, meaningful work, and not undermine his efforts. If they just throw him into unfocused micro-manged bureaucracy then it's just a waste.
It isn't clear to me that is it just comparing the javascript engines like for like. For all we know Google throws more operating system resources at processing the a single page than IE does natively. It may also be a battle of architecture (where each chrome page is a separate process with its own engine).
So it may not just be the skill of the engine developer, but competing architectures.
If Microsoft could throw away their old codebase and start fresh with all the hindsight we have now, including the Self papers of the mid 90s, I think they could do a much better job. Maybe not as good as Google, but decent. They could certainly match the Tracemonkey project.
Misleading yes but a valid point still. Microsoft could use WebKit in IE but instead their deep hatred and mistrust of open source software is standing in the way of providing their customers a better product.
To be fair it also may be the architecture that they've chosen for IE and windows. My understanding is that bits of IE form part of the windows O/S by it's use of COM: I'm guessing that Microsoft would like to honor a lot of the interfaces for these components.
MS makes money off desktop software, so putting a fast JS interpreter in IE would only benefit their competitors making web apps. It would border on negligence to their shareholders. Google, on the other hand, has economic incentive to promote faster browsers, as they're looking to make money off of web-based services.
Web-based apps are here to stay and over time will get bigger. By refusing to play the game, MS is burying its head in the sand and pretending the problem does not exist.
this is kind of unfair. the lessons microsoft and the development community have learned from working on browsers were all things that the Chrome team gets for free. they also get to start from scratch, whereas the IE team has to work on an outdated codebase and provide lots of legacy support.
Microsoft is free to rewrite IE from the ground up any time they'd like. The only legacy support they need is the same specifications that all other browsers are supposed to adhere to.
There are APIs for browser extension that Microsoft will want to continue to support as they are no doubt used by partners/corporations. It's not just about how well the web renders. The same way that firefox would be unwilling to start from scratch without providing support for the present plugin framework.
On the other, I think it's a stunning display of just how much muscle Google really has to flex (and perhaps more importantly, how much they're willing to flex).
It will be interesting to see how Google continues to treat companies it considers competitors over the next few years. The big guys like Microsoft will be able to handle it; but the smaller guys might have to start getting a little worried.
I would love to compete against Google. I would pretty much beg for the opportunity to do it. Other companies are too reactive; you do something, they do something, the entire competition is just a boring back-and-forth game of mostly minor incremental improvements of dubious benefit to the end users.
But Google? Now that would be fun. A lot of fun. If you let them get ahead of you even once, you've just dramatically decreased your odds of winning, and they have an awful lot of smart people working for them. You'd have to be more nimble, and you'd have to be able to guess at what they were going to do before they did it ... and then you'd have to do it better, and do it before they do.
Beating Microsoft is like playing chess in the park. Beating Google ... that's like playing speed chess with Big Blue. It would be a blast.
I hack on Arora a cross platform WebKit based web browser. And your right, I am having a blast. I have even contributed to Chromium so I know what I am up against :) Knowing that they have a team of really smart people working full time and Arora is a few devs in their free time really does make me think about what is important.
I did realize one thing though. While Chrome can never ship with adblock, Arora can, and 0.10.0 out next week will ship with it. :D
Have you considered how blocking ads is bad for the way online services are currently monetized? I'd rather have advertisers pay for my services than have to pay for it directly. If you don't like that trade-off, why not ask sites you use to offer an ad-free paid version? It sounds like you're trying to get the best of both worlds, and it isn't sustainable.
It isn't enabled by default so users still have to turn it on at this point. AdBlock was the highest voted feature that users wanted and was the biggest feature I could include that Chrome could not. From a bang for my time/effort AdBlock was a winner.
I'm no longer as fond of Firefox as I once was. It has persistent memory issues, and is getting more and more bloaty. But, it's also got AB+, and that for a lot of people is a killer feature.
I think there are a bunch of sound counter-arguments to the people saying that AB+ will kill their websites: one, users are sick of websites with more advertising-related content than actual content; two, fake virus alerts and other related ads are now one of the most common vectors for new spyware and malware; three, the ad revenue model is already starting to fall over due to over saturation, and maybe it's time that website owners start looking for a better way to do it.
So, thank you. I'll be trying out the next release of Arora.
Passive AdBlock use is a very implicit, inconsiderate way of telling people to get a new business model. How about, "Hey, the ads on your site are annoying, and I think I almost got a virus from one. I suggest you fix it. In the meantime, I'm using AdBlock on your site."
There are perfectly good reasons to block ads, but far too many people take a brute force approach to it that will result in good sites going away, or charging users instead of advertisers. I think both of those outcomes suck. If you want to block ads, I'm sure you can spare 30 seconds to let people know why and how to get you to stop.
I don't think this is flexing a lot of muscle, though I could be wrong of course. The VM was written by one very talented, experienced guy and his assistant. Let's say it cost $200k. The plugin probably cost a trivial amount (there is for example an open-source IE plugin for Firefox that is pretty small sourcecode-wise: http://ietab.mozdev.org/). So all in all a small bet with a large potential upside.
I don't think the technical aspect of what they've done is flexing muscle. It's interesting for sure, but in the end, it's a plugin (a very cool one, yes).
The muscle flexing is that they're willing to subvert a competitors product instead of just promoting their own product as a superior offering.
Can you imagine the outrage if Microsoft had done this and offered a plugin to Chrome or Firefox that would make them render things just like IE does? If you're having trouble imagining such a scenario, think back to the outcry when people discovered the .Net plugin for Firefox, and that wasn't nearly as invasive.
Google is drawing a line in the sand that essentially says "You can do things your way or our way, but if you don't like our way we'll find a way to make your users do it our way."
To me that seems like a pretty serious escalation.
Nah, this is no worse than making IE skins for Firefox, which have existed for some time. The .NET Firefox plugin was significantly more invasive, since it was automatically installed without user consent. As long as this plugin is voluntary, there's nothing serious about it imo -- it's more a fun prank that people may have developed with or without Google's help.
This is just speculation, but I think it's possible somebody just did this in their free time without much strategic planning, like so many projects at Google (you know, that 20% free time at Google).
I have some idea of how easy it is to embed IE into apps, so I am going to assume it's at least the same order of easiness to embed Chrome into IE. I think it's fairly trivial, though perhaps unpleasant work.
I tried looking at the source code, but I don't use GIT so I gave up. It's part of Chromium: http://code.google.com/chromium/
EDIT: After considering your profile page I fold :)
Again, I don't think that this is a technical thing; but a corporate strategy thing. We didn't hear about this from some developer's blog; we heard about it from an "official" Google blog. At some point, this thing has been blessed for release under the Google umbrella.
Anyone have real-life applications of this benchmark? Similar to 3D-gaming and CPU/GPU benchmarking in general the results (and winners) depend highly on the parameters of the test. Although at the end of the day nothing matters except real world performance.
Can someone point to an application where this performance improvement can be demonstrated?
It's only logical to conclude that Chrome standalone must run at least 8 times faster than Internet Explorer 8? Otherwise their gui would be much slower.
Why wasn't there a post about this before? This post must be complete bullshit.
Not really very relevant. Anyone who knows what a benchmark is is already running something other than IE. And who would install a Chrome plugin for IE when they could just as easily use Chrome?
If they don't realize what it actually is. There are a fair number of people out there who don't want to use another browser because it's confusing to them, or they like IE.
A website could just put a big "Speed up this application" button that installs the Chrome Frame plugin, and as far as the end-user is concerned, it's just like installing Flash - and they're still "using" internet explorer.
The target audience for this is not the people who are interested in Chrome; it's the people who aren't interested in moving away from IE.
But if that's the case, isn't that likely because they have internal applications that are IE only? It seems more like a marketing thing--Overcome people's inhibitions by letting them feel like they "only" installed a plugin and are still running IE.
Only websites with a specific meta-tag set are going to use the Chrome rendering engine, every other site will render using the default IE-rendering engine, including any internal IE-only applications.
Also, although it is hard to believe, some people actually think IE is a good browser. I don't understand how or why, but some people actually do.
If I were hosting a SaaS offering that was Javascript-heavy, I might deploy the site in such a way that IE users were prompted to download the Chrome plugin, thereby allowing my app to run at a tolerable speed.
We recently gave up on GWT at work for precisely the reason that IE6 and friends are waaaay too slow to do anything meaningful in GWT. Had we had this option we might have stuck with GWT instead of rewriting in Flex.
I think it's a matter of psychological leap (or the absense thereof).
Changing browsers requires aquiring new habbits. Most people don't like change. On the other hand, they're already familiar with the concept of a browser plugin thanks to Flash and Accrobat. The Google Chrome Frame can be understood as a speed booster and a bunch of additional features for IE rather than a new browser.
The key is that sites have to add a meta tag for the chrome browser to be used instead. It is opt-in only. This means that all those corporate sites that require IE won't be affected, but a user can still get the upgraded experience for sites that desire it.
there needs to be a way to force the rendering agent to be used, this would make the plugin agent much more viable instead of having to call cf: in front of a web address. I can see this being a good thing but also could be a bad thing, this doesn't fix the browser vulnerabilities that are still out there, merely your rendering engine. I would hate for people to be misled into thinking their version of internet explorer 6 is the latest and greatest because of the chrome rendering agent.
no, this is somewhat good news. IE was going to stick around for a long time anyways, because its the default windows browser and there will always be people who don't know better than to use the default.
this will just hasten the migration away. its probably not a huge win for anyone with respect to directly converting it into migrations to another browser, because the people who were clueless about switching probably still will be.
its just a really nice selling point when you have the ear of someone, trying to convert them.
Without knowing what site was used in the testing and in what manner (rendering speed, javascript speed, xml parsing speed ??) is the chrome plugin 8 times faster, I would rather keep my fingers crossed.
They used the SunSpider benchmark [1], which only tests javascript: "This benchmark tests the core JavaScript language only, not the DOM or other browser APIs."
Microsoft: "here's our best stab at what a great browser looks like - and it took us 14 years to get here".
Google: "we've been working on a codebase for 2 years. We decided to release it as a patch to your browser. Our brief foray into this area has improved your best effort by 10x".
Ouch!