> chrome has proved that continuous release cycle can produce the most stable and performant browser around.
I respectfully disgaree. New Chrome releases have broken both cosmetics (rounded corners, just as the CSS3 version was starting to take off) and basic functionality (H.264 support for the HTML5 video element is disappearing in about five weeks, even though it's there and basically working today, as far as I can tell for purely political reasons).
By the way, the word "performant" is real but does not mean what you think it does. Even if it did, I think you'd still be overstating the case. Chrome is probably faster than Firefox and IE in some respects today, particularly where JavaScript is concerned, but the gap has been closing for some time now. Once IE 9 and Firefox 4 are out, it looks unlikely that Chrome will still be significantly faster for typical real world browsing, which as I've noted previously, I consider to be more important than things like synthetic benchmarks or the speed of features no-one is using yet.
dom rendering speeds even up to ie8 lag far far behind what chrome and firefox have been capable off for years, everyone on the internet uses dom rendering and RIA's are not possible without being able to do fast dom modifications, most real world performance benchmarks have problems even running in ie because it is so slow.
h264 was a conscious choice, nothing "broke", I dont know which rounded corner issue you are referring to, it certainly isnt "broken" and remember you are referring to features which do not yet exist in any release of ie
You make a lot of generalisations about what "everyone" is doing and how "real world" things can't run on IE. If those claims were realistic then companies with widely used web apps, such as several of Google's more recent offerings, would not be able to support IE. However, in practice it appears that these web apps run on IE just fine, and IE continues to hold the lion's share of the browser market. I don't see how to reconcile that massive, worldwide data set with your version of how web development does (or, according to you, doesn't) work.
> h264 was a conscious choice, nothing "broke"
That depends on whether you're the start-up pushing a video-based service and trying to make use of the HTML5 video tag, doesn't it?
It used to work. In the near future, it won't. To most of us, that's a pretty good definition of "broken".
It can hardly be disputed that 1. pretty much everyone on the web is affected by dom rendering speed, 2. dom rendering is vastly slower in ie than chrome / firefox (for their related versions)
IE has a large market share, that has absolutely no correlation to its quality as a piece of software, web developers have to make their stuff work in ie because of its market share, lots of large web apps* either degrade or do not work in internet explorer even with the tremendous amount of extra effort required to develop for it comparatively to firefox + chrome
* Google Wave didnt support ie, offline gmail support is firefox only? google drawings didnt work in ie last time I checked, I am sure maps degrades in parts as well
This is the problem: the things you claim certainly can be disputed, not least because the second hasn't been true for several versions of IE. I know, the web apps I work on do a lot of this stuff, and there has been roughly an order of magnitude improvement in performance from IE6 to IE7 and then again from IE7 to IE8, which is now close enough to the other browsers that we don't notice the difference in day-to-day use.
I think you're making my point for me with the rest of your post: if you read it back, you'll see that you repeatedly contradict yourself about what does and doesn't work across different browsers, not just for IE but for Firefox vs. Chrome as well.
if you are going to dispute them dispute them with numbers, virtually every benchmark, independent or otherwise puts ie somewhere between bad and embarrassing
If you point is that we should just be happy with the web that ie6 supports then sure we wouldnt need to put up with fairly minor differences between firefox / chrome.
Personally I think we are still at an early stage of realising how computing can benefit our lives, for the last 5 years ie has been significantly slowing that progress down, hopefully recent signs continue to show that this progress will happen even if we need to drag ie along the way.
> if you are going to dispute them dispute them with numbers, virtually every benchmark, independent or otherwise puts ie somewhere between bad and embarrassing
1. Benchmarks aren't real world applications. I don't care if it takes 1s rather than 0.5s to run 1,000,000 operations, if all I ever need to do is run 5 of them.
2. The post you linked to is well over two years old. The entire web development world looked very different back then, and the possibilities of using JS and web apps for more serious work were in their infancy. At that point, Chrome did have a decisive advantage over all other browsers (not just IE) in JS execution speed, but of course that's just after Google had incorporated their new JS engine but before any of the more recent Firefox improvements and certainly not taking into account today's imminent releases like IE9.
3. I don't know what all those numbers are meant to mean, because there is no methodology information in the post you linked to. Charts and numbers without context prove nothing, rather like the "infographics" in the original article we're discussing here.
4. Even if you insist on using benchmarks, more recent ones (e.g., those highlighting jQuery 1.5 performance in some navigation features, published a few days ago) have IE9 still slower than other browsers, but more like a factor of 2-5 slower than the orders of magnitude it was just a few years ago. If they carry on improving at the rate they have been since IE7 -- keeping in mind that this seems to be one of their goals, and that the important context here is whether the IE team is following a healthy path for the future rather than whether they are better right now -- then we can expect IE10 to be directly competitive with the engines in other browsers.
As an aside, those few jQuery figures are a great example of the difference between real world usage and synthetic benchmarks. jQuery is a reasonably reputable bit of software, and the difference in performance in some basic document navigation tools -- hardly high-level code, but a step up from built-in JS/DOM functionality -- was as wide between jQuery 1.4.4 and jQuery 1.5 as it was between various pairs of browsers they tested.
> If you point is that we should just be happy with the web that ie6 supports
I've never said anything like that, anywhere in this discussion.
As it happens, I do think the latest HTML5/CSS3 features tend to be more about trend-setting than actually making web development more useful. If you want to make it useful, fire the guy at the W3C who thinks CSS shouldn't have tools like named constants and arithmetic, which web developers have been asking for since forever, and which everyone from programmers to those working in DTP considers entry level.
> Personally I think we are still at an early stage of realising how computing can benefit our lives, for the last 5 years ie has been significantly slowing that progress down,
How is that, exactly? IE has made a small impact in the speed of development of web sites. As we've seen, with competition to drive the market, Microsoft can still develop new browsers, they just chose not to for several years after IE6.
In any case, your point seems to assume that browsers are the only significant means for computers to benefit our lives, which is just crazy. If you want a disruptive technology, look at the iPhone: in just a few years, it has redefined the entire markets of both mobile computing and telecommunications, to the point where it's difficult to find any recent mobile phone that doesn't have a high-res touchscreen or any mobile platform that doesn't support installing custom software applications on the device, helping Apple climb back from relative obscurity in the PC world to one of the largest global brands in the process.
By the way, IE9 supports media queries -- a recent CSS feature that does have current widespread applicability -- just fine.
(Edit: Am I really being downvoted for arguing that qualitative factors and general policies/focus are more important than synthetic benchmarks, when the only quantitative data provided by the person criticising me for not citing hard numbers is one single blog post, from more than two years ago, containing a few numbers based on synthetic benchmarks, with no description of methodology to let us judge what if anything those numbers signify?)
> Benchmarks aren't real world applications. I don't care if it takes 1s rather than 0.5s to run 1,000,000 operations, if all I ever need to do is run 5 of them.
You are clearly forgetting where the web in general is going to. Browsers today serve as a wrapper to applications and NEED to be fast. Not a single web application today uses only 5 or 10 or 50 operations.
> IE has made a small impact in the speed of development of web sites.
IE has made BIG impact in the speed of development. The fact that it has not kept up with other browsers only made that worse.
Even IE8 rendering and its javascript engine have problems. Those are not present on modern browsers and it obviously impacts development because IE needs to have specific testing and specific coding (mostly workarounds).
Did you expect other browsers to be like IE, so that we could have a "standard" on not having standards?
> You are clearly forgetting where the web in general is going to.
No, I just take the view that the current trend of trying to treat a browser as an operating system is not going to last.
I think the web in the future will move back towards being a medium for presentation of information and some limited interaction, and things like native mobile apps and their app stores or Linux distros and their auto-installer tools will increasingly deal with what is sometimes done with "web apps" today. I think the important web technologies for the future will mostly be those that make it easier to present useful information, for example through more powerful presentation tools or providing useful context to web services so they can provide a more personal experience.
I already answered that question from someone else several hours before you posted it: no, I do not work for Microsoft, nor do I have any other connection to them that would bias my comments on this matter.
I simply disagree with what appears to be the majority view on HN in this case. Also, it doesn't help that a significant number of people seem to be responding to clichés that they assume I've repeated and not to what I actually wrote.
> Which version of Chrome broke the rounded corners?
They had rendering problems with several successive versions, as the "bug fixes" didn't for a while.
> What's its current market penetration?
That's not the point. They had a serious regression, which was pushed out to all Chrome users automatically because of their update policy. That one was cosmetic, but things like H.264 are breaking changes that remove whole chunks of functionality. The details don't really matter; the danger inherent in their process is the more serious problem.
> IE6 could learn from this.
Your example is somewhat ironic.
For one thing, IE6 is the bane of web developers everywhere because it didn't follow web standards -- not least because a lot of them hadn't been written yet. Sound familiar at all?
For another, IE6 is still used in a significant number of large organisations even today because those organisations value a stable, controllable platform as a foundation for their in-house developments more than they value the latest bells and whistles on the public web. Again, there's a lesson there for Chrome, Firefox, et al.
IE6 is still used in a significant number of large organisations because those large organisations have chosen to use implementations of fat expensive enterprise applications that have been customised sufficently to screw any chances of following the upgrade path for said fat expensive enterprise applications, meaning they'll only ever work properly in IE6.
At least that's the case where I work. Sigh.
> They had rendering problems with several successive versions, as the "bug fixes" didn't for a while.
Not so different than IE, right? It's really a speed of release issue, if I'm understanding you.
Well, it's a double-edged sword. Speedy releases mean speedy regressions and speedy bugfixes. Slow releases are the opposite.
> That's not the point
For me it is... but I think that ties in, below.
> significant number of large organisations even today because those organisations value a stable, controllable platform as a foundation for their in-house developments more than they value the latest bells and whistles on the public web.
We're definitely talking about some different markets here. The markets I work in simply don't target IE6 any longer. The baseline, if it includes IE at all--sometimes it's targeted at HTML-only specifically, is IE8 and whatever you can force IE7 to do.
When is IE6 going to cease to be the baseline in these other organizations? It's like its having a COBOL moment.
Hmmm... wiktionary includes a definition of "performant" as computer jargon meaning "capable of or characterized by an adequate or excellent level of performance or efficiency" [1]. I find it hard to disagree that this informal use has entered the computing lexicon (and don't personally object to it).
That "definition" includes eight distinct meanings. In time, any of them might be useful and the word might become an accepted part of the English language. For now, it just muddies the waters, when there are clear and unambiguous alternatives available. Perhaps that is why not one of the on- or off-line dictionaries I have just checked recognises any similar meaning of the word "performant".
This post has been brought to you by the "ironic parallels with the topic at hand" department.
> That "definition" includes eight distinct meanings.
"Performance" itself, and many other words, have lots of meanings [1].
> not one of the on- or off-line dictionaries I have just checked recognises any similar meaning of the word "performant"
True, but we have many terms in computing where our intended meaning is not captured by mainstream dictionaries.
> when there are clear and unambiguous alternatives available
Did you honestly find that usage unclear and ambiguous? The standard use of "performant" is so rare that there can be no doubt what the original poster meant. What are the alternative words you prefer? I think "performant" fills a gap.
> This post has been brought to you by the "ironic parallels with the topic at hand" department.
Nice observation! (perhaps not coincidently, I think I come down on the other side of the issue from you in both instances :)
I respectfully disgaree. New Chrome releases have broken both cosmetics (rounded corners, just as the CSS3 version was starting to take off) and basic functionality (H.264 support for the HTML5 video element is disappearing in about five weeks, even though it's there and basically working today, as far as I can tell for purely political reasons).
By the way, the word "performant" is real but does not mean what you think it does. Even if it did, I think you'd still be overstating the case. Chrome is probably faster than Firefox and IE in some respects today, particularly where JavaScript is concerned, but the gap has been closing for some time now. Once IE 9 and Firefox 4 are out, it looks unlikely that Chrome will still be significantly faster for typical real world browsing, which as I've noted previously, I consider to be more important than things like synthetic benchmarks or the speed of features no-one is using yet.