Chrome Browser Usage Artificially Boosted, Says Microsoft

bonch writes "Chrome was recently called the world's no.1 browser, but Microsoft is accusing the source, StatCounter, of using flawed methodology. When a user enters a search in Chrome, the browser preloads an invisible tab not shown to the user, and these were being counted by StatCounter. Net Applications, another usage tracking group, ignores these invisible tabs and reports IE at 54%, Firefox at 20.20%, and Chrome at 18.85%." Whereas the saturation of MSIE is totally organic, right?

Comments

I thought this was already refuted?

[Anonymous Coward]

StatCounter does not tally pre-loaded pages.

Re:I thought this was already refuted?

[MyLongNickName]

This might be what you are referring to:

"Last month, Net Applications began removing Chrome prerendered browsing traffic from its statistics, noting that “prerendering in February 2012 accounted for 4.3% of Chrome's daily unique visitors.” In doing so Net Applications became the first company to adjust its data reports for websites"

Re:I thought this was already refuted?

[MyLongNickName]

Ignore my sibling post, this is what I meant to grab:

"NOTE: StatCounter recently announced that they have updated their data as of May 1, 2012 to reflect prerendering in Chrome. However, there is no indication of either methodology or what percentage of Chrome share is being removed from StatCounter data."

Re:I thought this was already refuted?

[Moheeheeko]

Wait one fucking second

"bonch wites"

Theres our problem.

Re:I thought this was already refuted?

[PartOfElite]

Actually, many manufacturers pre-load Chrome now as Google pays them to. It's extra revenue for them, just like pre-loading Norton trial and so on.

On top of that Google also pays shareware and freeware authors to bundle Chrome in their installs, just like BonziBUDDY and all those other adware and spyware apps used to do.

Re:I thought this was already refuted?

[Calos]

Huh? The whole point of the GGP's post was that they recognize that there are other statistics services and to point out that those other services also claim that they ignore "Google's inflating tricks" - which, regardless, are not tricks meant to fool stats but to make things faster.

Re:I thought this was already refuted?

[hairyfeet]

Not to mention the thing that has most likely got MSFT worried which is that....NOBODY CARES, they really don't. This is one thing I have to give Moz credit for, because even though I no longer use their browser (I use a Chromium variant call Dragon) they were the ones that FINALLY got websites away from the "works best in IE" bullshit.

Now it doesn't really matter WHAT you use, its all the same. They all render the same pages, they all have roughly the same behavior, so the only ones that care about this little pissing contest is the corps themselves. as far as the users are concerned they honestly don't give a shit if what they are using is IE, Chrome, FF, dragon, QTWeb, Opera, whatever, it all "just works" and for that I say thank fucking God that it does.

Re:I thought this was already refuted?

[geirlk]

My wish is that MS would stop bundling their OS with a browser most use for downloading a browser, and actually bundled it with eg. Firefox instead.

Re:I thought this was already refuted?

[cpu6502]

Do you think Microsoft cares? They just want to spin the story to cover-up IE's downfall, and don't care if they have to LIE about StatCounter's methodology (claiming they count preloads, when they don't).

Re:I thought this was already refuted?

[ProbablyJoe]

So are MS just complaining about a problem which was already fixed before Chrome topped IE in their rankings? That's how it sounds at least.

From StatCounter's FAQ, and also noted on all of their graphs for this time period:

"Further to a significant number of user requests, we are now adjusting our browser stats to remove the effect of prerendering in Google Chrome. From 1 May 2012, prerendered pages (which are not actually viewed) are not included in our stats."

http://gs.statcounter.com/faq#prerendering

Re:I thought this was already refuted?

[gsnedders]

Except that's not the case. Go to any Google website in Opera: it looks different. Why? UA sniffing. Go to Facebook in Opera Mobile: it looks different. Why? UA sniffing.

In both cases Opera functions fine if you change the UA string. Sadly, evangelism isn't enough to fix everything.

Re:I thought this was already refuted?

[BZ]

The problem is that we're rapidly moving back to the "works best" bullshit, but now with "Chrome" or "WebKit" in place of "IE"...

Re:I thought this was already refuted?

[Scarred Intellect]

Not to mention the thing that has most likely got MSFT worried which is that....NOBODY CARES, they really don't. This is one thing I have to give Moz credit for, because even though I no longer use their browser (I use a Chromium variant call Dragon) they were the ones that FINALLY got websites away from the "works best in IE" bullshit.

Now it doesn't really matter WHAT you use, its all the same. They all render the same pages, they all have roughly the same behavior, so the only ones that care about this little pissing contest is the corps themselves. as far as the users are concerned they honestly don't give a shit if what they are using is IE, Chrome, FF, dragon, QTWeb, Opera, whatever, it all "just works" and for that I say thank fucking God that it does.

Opera and Safari don't work for many of my school's intranet functions. Opera doesn't work for the FAFSA and a number of other scholarship applications (To be fair, the FAFSA works, but it isn't supported). They don't work for many job applications I've done. I think VONAPP through the VA doesn't render correctly. Hell, even my slashdot journal editing doesn't render correctly in Firefox, Chrome, or Opera (had to use IEx64 for it to render correctly). A great majority of websites work independent of the browser, as they should, but some still don't, and unfortunately for me and every student in the country, many of these are from our school or the government.

Re:I thought this was already refuted?

[gx5000]

You're going to get flack no matter what you post here, but after twenty five years of IM/IT, I agree with you... Most people are too lazy or non tech educated to get a different browser anyways.... Take that plus the fact that Chrome plugs into all things Google and rips your privacy apart, nevermind the footprint too and I'll stick to Opera and IE, since that gives me speed and compatibility in that order. I can't count the times I've had people bitch because Firefox or Safari or etc wasn't brining up their code properly, but IE did....ugh.

Re:I thought this was already refuted?

[wvmarle]

Long time since I've seen that, and most sites where I see "works best" it's something like "Internet Explorer 5.5". Oh well.

I can't believe web sites nowadays can afford to have it work well in one browser, and not so well (missing bits/poor layout/whatever) in the rest. Because the most-used browser is only just over half of the users according to one set of statistics, and about a third of the users according to another set. So half or more of your users will see a degraded site.

Re:I thought this was already refuted?

[Kjella]

Opera doesn't work for the FAFSA and a number of other scholarship applications (To be fair, the FAFSA works, but it isn't supported)

Those are two completely different things if you ask me, particularly if IE, Chrome, Firefox and Safari is supported. Supported to me means tested (okay don't laugh), that someone actually fires up that browser and goes through the use cases. That takes time and money every time either the software is updated or the browser is updated. At some point you make a cut-off saying all the remaining browsers are on their own, they might work but we're not going to guarantee they work. As long as that cut-off list is at least 3 different browsers I'd say they're far, far above the pack when it comes to support. If it works on everything else but Opera, chances are pretty good it's a bug in Opera and that's where you should direct the bug report, not the website.

Re:I thought this was already refuted?

[BZ]

> Long time since I've seen that

Take a look at http://getcrackin.angrybirds.com/

If it were just "works best", that would be one thing, but the new trend is a real throwback to the days of "we just won't let you in if you're not using the one browser we approve of"

> Because the most-used browser is only just over
> half of the users

True on desktop. Not so much on mobile. And oddly enough sites that do UA sniffing and serve different content to "mobile" browsers have all sorts of WebKit-only stuff going on.

But even on desktop, people seem pretty happy to have it look broken in whatever browser they don't happen to be using.

And we're not talking just small sites. Google has had several instances recently where updates to things like Gmail got rolled out that were apparently not tested in any non-Chrome browsers or something, since they only worked in Chrome. At least Google considers that sort of thing a bug and fixes it quickly....

Re:I thought this was already refuted?

[recoiledsnake]

Looks like you're the one spinning the story. The Microsoft blog post linked is from March 18 2012.

http://windowsteamblog.com/ie/b/ie/archive/2012/03/18/understanding-browser-usage-share-data.aspx

Re:I thought this was already refuted?

[shutdown -p now]

Now it doesn't really matter WHAT you use, its all the same.

-webkit-bells-and-whistles: 100%

Re:I thought this was already refuted?

[MightyYar]

But this is inevitable... you will always have a feature-rich website that you test on a handful of "supported" configurations. If the UA doesn't match the supported configuration, you fall-back to a safe version of the site. You can't possibly test every configuration, and even if you could it wouldn't make any financial sense to do so.

I think it is unrealistic to ask, for instance, Google to just serve up the same page to everyone and let the non-conforming browsers fall by the wayside. They don't want to turn away advertisement targets.

Re:I thought this was already refuted?

[Asksa]

The blog post is two months old. It wasn't fixed back then.

Re:I thought this was already refuted?

[JustSomeProgrammer]

I think it completely depends on which browser you started coding against. If you do your main testing against IE it might look wrong in Firefox until you work out the differences. If you started in Firefox IE will look weird.

Re:I thought this was already refuted?

[Lord_Jeremy]

Hah, I wish! I'm still designing pages that are compatible with IE5.5+ which means accounting for all sorts of annoying css render bugs. Even recent versions of IE exhibit things like the float overflow drop bugs. The IE developers seem to have this terrible notion that no matter what the CSS standard actually says, web designers and other web browsers are supposed to follow their lead as to how certain properties behave. That's what it seems like at least, considering the number of layout bugs that have been in every IE since around 5 to the latest versions.

Re:I thought this was already refuted?

[chrb]

StatCounter recently announced that they have updated their data as of May 1, 2012 to reflect prerendering in Chrome. However, there is no indication of either methodology

They state the methodology in their FAQ:

BROWSERS: Do you adjust your browser stats for prerendering/pre-loading?
Two browsers are affected by preview-type requests - Chrome and Safari.
Chrome
Further to a significant number of user requests, we are now adjusting our browser stats to remove the effect of prerendering in Google Chrome. From 1 May 2012, prerendered pages (which are not actually viewed) are not included in our stats.

Some points to note:
Prerendering was announced by Chrome in June 2011. This change did not have any significant impact on our stats.
Chrome is currently allowing the detection of prerendering behavior via its Page Visibility API.
Google specifically states: "
Important: This is an experimental API and may change-or even be removed-in the future, especially as the Page Visibility API standard, which is an early draft, evolves."

This means that in the future it may not be possible to track/remove the effect of prerendering on Chrome. If other browsers adopt prerendering then it may not be possible to track/remove the effect of prerendering on those browsers. In that case, the fairest solution would be to include all page views (prerendered or not) for all browsers rather than only excluding prerendering in Chrome. That scenario would require us to revisit this methodology change in the future.

We publish a graph showing total prerendered page views tracked in Chrome, together with the portion of prerendered pages which are not actually viewed by the end user. The prerendered pages (which are not actually viewed) are removed from our stats. For May 2012, the percentage of prerendered pages (not viewed) in Chrome is approximately 1.3%. Note that this change has not had any significant impact on our browser stats. This is due to our use of page views to track browser usage - page views are less susceptible to influence by prerendering than unique visitors.

Safari

The Top Sites feature in Safari shows preview thumbnails of frequently visited sites. These preview thumbnails are refreshed by Safari periodically. Unfortunately, it is not possible to exclude these previews from being tracked. To get a bit technical, this is because the "X-Purpose: preview" header is only sent with the request for the base page. The header is not sent as part of requests for images, CSS or JavaScript that have to be downloaded and executed as part of the Top Sites preview. With online web analytics (as provided by StatCounter) the relevant header information is not passed so these preview requests can't be detected and therefore can't be removed. Ideally Safari will change this to ensure to send the "X-Purpose: preview" header with all Top Sites HTTP requests, however this is not the case at present.

Interesting that Safari is still being over-counted though.

Re:I thought this was already refuted?

[Tough Love]

You know what's really annoying? The fact that all the Webit browsers identify themselves as "Apple Webkit" when it's really "KHTML", a product of KDE volunteers.

Re:I thought this was already refuted?

[MightyYar]

I was responding to the "go to any Google website in Opera" comment by gsnedders. I don't doubt there are demo or game websites that work in only one browser or another.

Re:I thought this was already refuted?

[Anonymous Coward]

But this is inevitable... you will always have a feature-rich website that you test on a handful of "supported" configurations. If the UA doesn't match the supported configuration, you fall-back to a safe version of the site. You can't possibly test every configuration, and even if you could it wouldn't make any financial sense to do so.

I think it is unrealistic to ask, for instance, Google to just serve up the same page to everyone and let the non-conforming browsers fall by the wayside. They don't want to turn away advertisement targets.

what you describe is how things were done 5-10 years ago - today you are supposed to test directly for available features, not for UA strings.

Re:I thought this was already refuted?

[CowTipperGore]

Indeed. I ran Firefox for years but had to keep IE around (using the IETab add-in) for particular sites. When I moved to Comodo Dragon a few months ago, I put it on my work laptop, our home laptop, our two netbooks, and our old XP desktop. I haven't needed or used Firefox or IE since. Sure, a few sites bitch that I'm on an "unsupported browser" but they work fine, and a rare few sites screw up the page layout. But, unlike using Firefox five years ago, I've yet to find anything that simply doesn't work or that absolutely requires IE. Progress baby.

Re:I thought this was already refuted?

[MightyYar]

You still have to do QA, and the safest thing to do after deciding what configurations you run QA on is to default to a lowest-common-denominator version of your site. Remember that Google in particular would rather you see a gimped version of their search site than have some crazy error, crash, or display defect.

Re:I thought this was already refuted?

[Anonymous Coward]

Working only in one browser is fine if you only are using tech from that browser (although you really should offer at least a degraded experience).
Also is fine if you're too bloody lazy to add the alternative prefixes for other browsers, although that is kinda lame.
Heck. Also ok to say "Use Chrome" even though Firefox works fine, or sometimes even better. That's kind of lame, but ok.

What's evil is this blocking of other browsers and saying "download chrome" even though there is nothing wrong with the other browsers apart from you blocking their UA.
This is Chrome specific behaviour. This as well as the above behaviours are common on their demos.
Mozilla has occasionally made demos that used something like the sound specification they were proposing for the web. Those demos *never* blocked other browsers though, and in fact worked fine in Chrome, just w/o same performance or sound synchronisation.

Ditto Chrome's app store. Support for other browsers is crap in terms of outright blocks, use of chrome specific stuff all over the place, and no attempt at cross-browser at all.
Compare that with Mozilla's app store.
https://wiki.mozilla.org/Marketplace/Compatibility

Where compatibility is actually a priority.

Basically, Google *is* breaking the web which is why this is relevant to the parent post.
I fear for the state of the web if they get supremacy. I don't see them being any better than Apple with their closed demos or Microsoft with their proprietary apps.

Re:I thought this was already refuted?

[s.petry]

One point I believe was being made is that average users in most cases don't know about different browsers. This would be the average person that goes to Best Buy or and just buys a computer for home use, which is most people. Add to this the incredible amount of IT departments that have IE as the only browser installed for their consumers. Average consumers don't know any better, and I'm sure that many companies prefer it this way. It is cheaper in IT providers to keep the default browser, and no extra effort is required to maintain IE.

I think the big push to ".net" has been a big attempt at re-initiating the old browser wars, and Microsoft has had a lot of success with this.

Because most people don't know better (family), I teach them "www.getfirefox.com" to at least get them off of IE. It's simple for them to use and set up and takes no effort outside of explaining that a "Bookmark" is the same as a "Favorite".

Re:I thought this was already refuted?

[cpu6502]

>>>March 18 2012.

Oh.
Then why the hell is this a story? It never should have been posted to slashdot. It's two month-old news.

Re:I thought this was already refuted?

[LDAPMAN]

That top-sites preview page is rarely loaded, especially if you have changed your home page. I would not expect it to have much effect on the stats. Chrome pre-rendering happens on virtually all pages.

Re:I thought this was already refuted?

[Kalriath]

Because that's what Slashdot does... Give you news that mattered - months ago.

Re:I thought this was already refuted?

[hairyfeet]

Hey another Dragon user! Nice! Have you tried readability yet? Great extension, takes all those damned websites that spread the article over a dozen pages and just makes it one long page. If you use it with adblock it makes it like the old days, when you could actually read on the web, it even lets you set font size! Give it a spin if you haven't already, you won't regret it.

But I have to agree, since switching I haven't ran into any "IE" only sites, and the only site they gave as an example i couldn't get to load completely was some angry bird nut promo thing...like I really care about the angry birds pushing nuts.

Re:I thought this was already refuted?

[fiziko]

Not to mention xkcd on April 1, 2012...

Re:I thought this was already refuted?

[omfgnosis]

There are much better ways to provide fallbacks to feature-rich web experiences than UA-sniffing. Using proper feature detection in JS and some basic (really!) understanding of how to provide basic styles behind new CSS features, you can provide the best experience for the best user agents without excluding anyone, and without the burden of testing every possible configuration. These are code practices that have been possible since the first "browser wars" ended, and have been advocated heavily for at least a decade.

I've had great success with progressive enhancement, even for pretty involved work. I recently completed an rather large (for me) project with a lot of bells and whistles—including substantial use of APIs that were only introduced in the last couple years, as well as a really nice interface that uses no images and is perfectly usable on browsers with poor graphical capabilities—and the cross-browser phase of the project was so minimal as to be laughable. Our small firm, of course, only commits to supporting certain configurations for our client, but being a good guy I checked a bunch of other configurations to see what the damage was. Apart from IE 6 and 7 (which we no longer support), there was literally no damage, at all. And that includes testing in Opera (which we don't even discuss with clients, most of whom have never even heard of it) as well as versions of Firefox that are so old that they no longer register on stats anymore. And IE 6 and 7 support took a matter of a couple hours, so I went ahead and did it. And there isn't a bit of browser branching in the code besides the use of a few classes output with IE conditional comments.

Do I (or my small firm) know something Google (or Apple, who are at least equally guilty of overly restrictive UA-sniffing) don't? Of course not. Presumably, they maintain an internal list of supported configurations and block everything else as a matter of managing cost. But I seriously doubt that the list is a product of extensive testing which revealed bugs in the otherwise-perfectly-capable configurations that they block. They simply don't check, and operate from a fundamentally poor approach.

The best part is that, if you begin from a progressive enhancement approach—with a little bit of foresight and as much experience as anyone else doing anything so "feature-rich"—none of this is difficult. Cross-browser has only gotten easier since we adopted this approach, and the number of target configurations has only grown.

I'm sorry (I really am) if this comes off as preachy, but I am astonished to see, in 2012, arguments in favor of UA-sniffing. With all that's wrong in the standards, in the history of the standards, and in the massively broken implementations over the years, the solution has been clear for a long time. Just because a behemoth like Google doesn't get it right doesn't mean it's wrong.

Re:I thought this was already refuted?

[omfgnosis]

The safest thing to do is to test for the features you want, fall back for the features you need, and let the browser do the rest of the work for you. It's been possible to do this since Netscape 4 went out of style, and Google in particular isn't doing anything so demanding with their search site that they'll find the kinds of edge cases that actually expose the broken edges of otherwise solid implementations.

In a perfect world, developers would be able to enforce upgrades to any rendering engine to ensure the best experience. But for years and years there have been APIs designed, albeit poorly overall, well to handle backwards compatibility as well as forward compatibility—and for the few exceptions, IE has conditional comments.

There are some features that absolutely require a new and working implementation, and for those features there are fallback capabilities as well. It takes no UA-sniffing at all to write <canvas>Your browser doesn't support my drawing app</canvas>, nor <video><source src="foo.mp4">Your browser doesn't support AVC</video> (apologies if this isn't the correct fallback syntax, it's one I haven't spent much time with, but I suspect it's correct and it's pretty intuitive if you know how to code fallbacks for the web).

Re:I thought this was already refuted?

[hairyfeet]

Chrome might but Comodo Dragon don't, it says in its UI string "AppleWebKit/536.5 (KHTML, like Gecko) " and then it has Chrome Safari and Comodo Dragon. so maybe they are just covering all the bases but at least one does give KHTML credit.

Re:I thought this was already refuted?

[MightyYar]

Your browser doesn't support my drawing app

Right, but don't you have a marketing department? They would never want anything like that displayed on the page.

Re:I thought this was already refuted?

[MightyYar]

I notice you didn't mention anything about automated tests... add that to your requirement :)

Re:I thought this was already refuted?

[MyLongNickName]

I find it funny that some dolt modded this troll. The post itself immediately got modded back up, so the person wasted a point.

Re:I thought this was already refuted?

[tjhart85]

Chromium is fully open source & has most of the benefits of Chrome. If you're paranoid enough to not use Chrome because of privacy reasons, just use Chromium instead.

Re:I thought this was already refuted?

[omfgnosis]

Why? Sure, in some cases, depending on complexity and the actual testing requirements, it may be a worthwhile investment, but I don't think it is relevant to the question of progressive enhancement as a development approach, and certainly wouldn't be relevant to the benefits of progressive enhancement on the configurations you don't even support. The question is, for those configurations, do you employ known-good development practices that will support them to the extent of their capabilities with or without testing, or do you arbitrarily block them from the functionality and present them with an ugly notice saying they need to choose a platform that you approve? Automated testing doesn't benefit a platform you don't test in the first place.

Re:I thought this was already refuted?

[omfgnosis]

So what's the alternative? Recalling that you promoted User Agent sniffing and arbitrarily blocking untested clients, I don't see the fallback notice as fundamentally different from a marketing perspective, except that UA-sniffing will cause more false positives and present your block message to more users needlessly. The alternative, so far as I'm aware, is to develop two apps in tandem: a canvas/JS app and a Flash or Silverlight fallback. Marketing (if it's hopelessly large and powerful) will probably go for that, but it could mean potentially twice the cost or more.

Let me reiterate: There are some features that absolutely require a new and working implementation, and for those features there are fallback capabilities as well. Are you just objecting to the language I used in my pseudocode fallback? Or are you really saying that the approach you promoted—UA-sniffing—would provide a better user experience than the built-in fallback capabilities of canvas and would better appease a marketing department that doesn't understand these new APIs?

Re:I thought this was already refuted?

[MightyYar]

Or are you really saying that the approach you promoted—UA-sniffing—would provide a better user experience than the built-in fallback capabilities of canvas and would better appease a marketing department that doesn't understand these new APIs?

No, I'm saying that it is reasonable to only put out a "rich" version of your site for tested configurations. What is the point of having unit tests, regression tests, and such in place if you are going to then allow configurations that are not part of your test suite? At that point, the safest thing to do is put out the simplest version of your site that marketing will accept and hope for the best.

Re:I thought this was already refuted?

[omfgnosis]

No, I'm saying that it is reasonable to only put out a "rich" version of your site for tested configurations.

And you haven't advanced any argument for that other than say-so! I've advanced an argument based on evidence—years and years of APIs designed specifically for this purpose, and years and years of a safe environment of "unknown" configurations in which to deploy.

What is the point of having unit tests, regression tests, and such in place if you are going to then allow configurations that are not part of your test suite?

Your target environments pass the tests. I don't see why this is a point of debate. The benefit of progressive enhancement is that it's quite likely the environments you didn't test would also pass.

At that point, the safest thing to do is put out the simplest version of your site that marketing will accept and hope for the best.

No, it's really not. You haven't addressed progressive enhancement at all, why don't you try? The safest thing to do is to test for the features you use, and provide the "simplest version" of those specific features when those features are unavailable. Why should Opera users be arbitrarily restricted from using your app? There isn't any reason. Opera is perfectly capable of running it. You just didn't test it. If a specific API is unavailable, you have a "simpler version" implemented and fall back to that for that API. The web isn't all or nothing, and with a moving target API it never will be.

You can certainly provide a notice that the environment is untested and some features may not be available, but why just block usage? If you're writing good, progressively enhanced code in the first place, there's no benefit to then drawing completely ignorant lines in the sand as to the capabilities of environments you haven't even tested.

Re:I thought this was already refuted?

[MightyYar]

Automated testing doesn't benefit a platform you don't test in the first place.

You are correct - but there is less risk in the first place if you offer those untested configurations the simplest site that is acceptable to marketing.

It's all about tradeoffs... If we return to the Google example, they use both strategies. On their main search page, the parts that require a modern web browser are mostly fluff - the auto-search-as-you-type, the animated logo, the Google+ integration, the preview sidebar... Very little actual functionality is lost by just showing a simple, old-fashioned Google search page to unknown clients.

Now surf on over to Google Docs and try to use that with an unsupported browser. First it tests for JavaScript and warns you about that. Then it does some kind of UA check, directs you to a link showing what platforms are supported, but tells you that you can still use the page by adding ?browserok=true to the url (along with a shortcut link that does this for you). At this point, it gives it the ol' college try.

I think that is a very reasonable balance of the two methods, because it lets a knowledgeable user try without resorting to UA spoofing, but lets you know that your configuration is untested and the functionality is not guaranteed. I contend that their search page is simple enough that it wouldn't make any sense to burden the user with these extra steps.

Re:I thought this was already refuted?

[MightyYar]

You haven't addressed progressive enhancement at all, why don't you try?

I see it as an additional risk and an added complication. Google uses it (after first warning you that you are unsupported) with their Docs service, but not for their home page - which benefits only a little from the modern features anyway.

Try visiting Google Docs in an unsupported browser and see what I mean. Can you honestly say that the meager difference between unsupported Google search and supported Google search merits this kind of system?

Re:I thought this was already refuted?

[omfgnosis]

You are correct - but there is less risk in the first place if you offer those untested configurations the simplest site that is acceptable to marketing.

No, there isn't! There's no risk at all of providing more features to an unsupported environment.

It's all about tradeoffs... If we return to the Google example, they use both strategies. On their main search page, the parts that require a modern web browser are mostly fluff - the auto-search-as-you-type, the animated logo, the Google+ integration, the preview sidebar... Very little actual functionality is lost by just showing a simple, old-fashioned Google search page to unknown clients.

Sure, it's a contrived example. But even so, Opera is capable of all of those features, and if the code is written properly in the first place it would require no extra work on Google's part. Actively blocking access to those features requires more work.

Now surf on over to Google Docs and try to use that with an unsupported browser. First it tests for JavaScript and warns you about that. Then it does some kind of UA check, directs you to a link showing what platforms are supported, but tells you that you can still use the page by adding ?browserok=true to the url (along with a shortcut link that does this for you). At this point, it gives it the ol' college try.

So, imagine a scenario where your browser just adds "&browserok=true" to every Google URL. How does this harm or increase risk to Google?

I think that is a very reasonable balance of the two methods, because it lets a knowledgeable user try without resorting to UA spoofing, but lets you know that your configuration is untested and the functionality is not guaranteed. I contend that their search page is simple enough that it wouldn't make any sense to burden the user with these extra steps.

Why not just provide the notice and be done with it? Why require any step from the user?

Re:I thought this was already refuted?

[omfgnosis]

I see it as an additional risk and an added complication.

I've already addressed risk in another post, but I'll ask again: what risk? You already don't support the configuration, what is the risk in letting the user's browser use the features it supports?

As far as complication, I can assure you that you're wrong. Progressive enhancement greatly simplifies web development. It simplifies testing. It simplifies cross-browser fixes. It simplifies your code, and ensures a greater degree of correctness from the outset.

Can you honestly say that the meager difference between unsupported Google search and supported Google search merits this kind of system?

Yes, absolutely. The difference between the two isn't at issue, what's at issue is the harmful approach of wrongheaded development approaches:whitelisting browsers based on name rather than safely using features as they are available. Even the simplest site merits progressive enhancement, because it's the correct approach to arrive at a universally usable site without making arbitrary and ignorant guesses about things you can't possibly know.

Keep in mind that there are new browsers coming out pretty frequently, and being prepared for them is better than not. Progressive enhancement provides that kind of forward-compatibility. UA-sniffing by definition does not.

Re:I thought this was already refuted?

[lsatenstein]

Due to FFox anomalies, I do half my browsing with IE and the other with FF. IE leaves me more room on my netbook screen for text. With FF, I have to hit the F11 function key in order to see what I can with IE as is.

I do like FF for the multilanguage support.

Re:I thought this was already refuted?

[MightyYar]

what is the risk in letting the user's browser use the features it supports?

The risk is that your site will fail in a way that you did not anticipate.

As far as complication, I can assure you that you're wrong. Progressive enhancement greatly simplifies web development. It simplifies testing. It simplifies cross-browser fixes. It simplifies your code, and ensures a greater degree of correctness from the outset.

You can't possibly state that without knowing the starting condition of the code and the ultimate complexity of the site you are coding.

Yes, absolutely.

Really? It can't ever be cheaper to implement a UA check than a feature check?

Progressive enhancement provides that kind of forward-compatibility. UA-sniffing by definition does not.

I'm gonna call you on that - the simple search site that Google serves up to unsupported UAs is probably going to render on any decent browser made from this day forward, even if they never do another ounce of coding.

Re:I thought this was already refuted?

[MightyYar]

There's no risk at all of providing more features to an unsupported environment.

Sure there is - your site may fail in a way that you didn't anticipate. Without a test, you would never know this.

Sure, it's a contrived example.

Well, that's my point :) In real-world examples, there is always a tradeoff. You can't just say "always use feature detection" because that is not always necessary... there are cases where the simple site is not significantly less functional than the full-featured site, and limiting your test configurations at the expense of browser support becomes economically viable.

Actively blocking access to those features requires more work.

I disagree. Screening for UA is very, very simple. I'd hesitate to call it "work" except that you will probably get paid to do it. Further, this solution works with existing code. If someone handed me a pile of old code and told me to make it work with IE6-9, Firefox 3-11, and the latest Chrome, I'd probably jump to the UA solution initially. Feature detection is not something that would easily bolt-on to legacy code. If I were writing from scratch, I would probably use feature detection, but then I'd still do a UA check against my test configurations and warn the user if they are unsupported. So either way, I'd have to do the UA check.

And that's if marketing lets me ugly-up their site with a warning. If they don't, I might still have to fall back to a basic version of the site to be conservative.

So, imagine a scenario where your browser just adds "&browserok=true" to every Google URL. How does this harm or increase risk to Google?

I imagine it would increase their tech support costs.

Why not just provide the notice and be done with it?

That's just a design choice. They happened to choose a hyperlink that brings you to the unsupported page. They could have just as easily done it with a pop-up, alert, modal, or some other user notification. Any good solution would still require the user to click to acknowledge the warning - you don't want to just pop up a box for a few seconds and auto-hide it, and there's no reason to ugly-up the page with a permanent warning.

Re:I thought this was already refuted?

[omfgnosis]

The risk is that your site will fail in a way that you did not anticipate.

The hell it will! It's quite obvious you don't know the first thing about progressive enhancement. The entire purpose is to anticipate failures, and to accommodate them by falling back to a simpler version; this is effectively what you would expect in your supported environments as well. In some cases, there simply is no simpler version (the example I used is canvas), in which case the API correctly anticipates this and allows a fallback message—the same fucking message you would present with an arbitrary and overly broad UA-based block. In the absolute worst case scenario, you get the same outcome you're promoting. Everything else is improved functionality for every user of a modern browser you didn't test. There is literally no risk in this approach that isn't inherent in web development, and there's substantially less risk of putting off customers/clients/users than the use of arbitrary UA-sniffing.

You can't possibly state that without knowing the starting condition of the code and the ultimate complexity of the site you are coding.

Yes, I can! I know it from experience, both from inheriting existing code bases and from building extremely complex web applications. The existing code bases that depend on arbitrary UA-sniffing are by definition incomplete for future browser releases, and require either tedious and labor-intensive maintenance or a complete rewrite as they age; the existing code bases that depend on feature detection effectively work the same or improve with future browser releases, and require little to no maintenance for existing features; the complex web applications I built in the past that depended on UA-sniffing or other naive browser-based branching were much the same as inherited UA-sniffing code bases; the complex web applications I built with progressive enhancement required none of this addition complexity or effort.

Sure, that's all anecdotal. Okay, let's just look at the code and its implications. Understand that I'm writing simplified pseudo-code for the sake of brevity and digestibility.

// Here we determine whether fullscreen can be entered, if so we enter it, else we provide a fallback
if(someNode.requestFullScreen) {
someNode.requestFullScreen();
} else { // Fail gracefully
}

// Here, knowing that IE versions through 9 don't support fullscreen, so we detect IE version and attempt to enter fullscreen
if(!$.browser.ie || $.browser.ie >= 10) {
someNode.requestFullScreen();
} else { // Fail arbitrarily
}

(Note that the second example is abusing pseudocode, as presumably your UA-sniff would be outside the script entirely and restrict access to the entire script. But it's fundamentally the same approach.)

Do you see the problem? Obviously, being simplified, the actual list of arbitrary browsers (whether a whitelist or blacklist) would be much more ridiculous and arbitrary and error-prone. But, to simplify, what if IE10 doesn't support fullscreen? Runtime error, failure to execute code below. Well, suppose we exclude IE10 as well. We're then adding the additional cost and labor (including tracking this as a task to do, and the mental cost of revisiting a particular implementation detail potentially months later) of going back and correcting if IE10 does support fullscreen, or even more likely simply forgetting to do so and leaving all IE10 users in the lurch, potentially some dozens of percentage of your users.

The first example simply lacks any of this cost or potential for failure, and is much simpler to implement.

There's a third approach—combining the two—and I'll give you the benefit of the doubt and assume this is what you'd a

Re:I thought this was already refuted?

[omfgnosis]

I replied to some of the points in this post in the other post, so I'll refer to that response rather than repeating myself. I am not dodging, just adhering to DRY principles!

Well, that's my point

And you conveniently omitted the context where I explained the absurdity of your point. Being a contrived example, I pointed out that yes, while "very little functionality is lost", there is still no benefit to arbitrarily excluding clients which are perfectly capable of that functionality.

You can't just say "always use feature detection" because that is not always necessary

Yes, I can, and yes, it is. You still have to do the feature detection to support your known environments. There's no reason to do additional work for no benefit.

there are cases where the simple site is not significantly less functional than the full-featured site, and limiting your test configurations at the expense of browser support becomes economically viable.

I'm not advocating adding test configurations, so I don't know what you're talking about. Using proper development techniques, there's no reason to add test configurations.

I disagree. Screening for UA is very, very simple. I'd hesitate to call it "work" except that you will probably get paid to do it.

1 > 0. It's not a matter of fucking opinion.

If someone handed me a pile of old code and told me to make it work with IE6-9, Firefox 3-11, and the latest Chrome, I'd probably jump to the UA solution initially.

Then they should hand someone else the code.

Feature detection is not something that would easily bolt-on to legacy code.

You're right. That's one huge reason that UA-sniffing as a legacy is so fucking harmful. Properly written code from 10 years ago (that is, using feature detection) will not have this trouble.

If I were writing from scratch, I would probably use feature detection, but then I'd still do a UA check against my test configurations and warn the user if they are unsupported. So either way, I'd have to do the UA check.

Write code which properly requests client features, and warn for unsupported configurations? Sounds like exactly what I've been suggesting this whole time. Gee whiz!

And that's if marketing lets me ugly-up their site with a warning. If they don't, I might still have to fall back to a basic version of the site to be conservative.

If your code is written properly, the warning is excessive insurance, and the fallback will happen automatically. Marketing would surely prefer better functionality than worse functionality, yes? The warning can be in your terms, on your support page, or any discreet place you wish. And it doesn't require a single line of UA-sniffing code, it can be in plain HTML.

I imagine it would increase their tech support costs.

How the fuck would it do that? They don't bloody support Opera at all.

That's just a design choice.

Yes, a code design choice! It's a code design that does not arbitrarily block unknown configs that are perfectly capable of identifying the features they support.

Any good solution would still require the user to click to acknowledge the warning - you don't want to just pop up a box for a few seconds and auto-hide it, and there's no reason to ugly-up the page with a permanent warning.

A better solution would be to provide an unobtrusive but present-until-acknowledged warning that goes away until its acknowledgment cookie expires. An even better solution would be to write safe code in the first place, providing a notice in some discreet location as extraneous insurance if you're paranoid.

Re:I thought this was already refuted?

[deppman]

I am not trolling, just an observation: I very much care what browser my customers use, and they do too.

I care because I am a professional web application developer, and am sick to death of writing two sets of code: One for standard compliant browsers, and one for IE. We have to do this because until recently the leaders at MS have clung hopelessly to the belief that they set the standards regardless of the written specs.

The only reason IE is getting better is because they are getting pummeled in the market. We all saw what happened when IE6 was the defacto standard - 10 years of stagnation before the first significantly standard compliant browser from MS (IE9) was released. For example, their much touted 'native SVG support' came only 10 years after KHTML (which became WebKit) and 6 years after Firefox (v1.5). If it weren't for the can of whoop-ass opened on them, I highly doubt this or other features would have seen the light of day, since their proprietary Sliverlight was supposed to be the future of the web.

My customers care because we develop for webkit and firefox first. This is not religion, just good business sense - these execution environments represents the broadest swath of desktop and mobile users. We degrade our applications the best we can for IE. In the worst cases, I have asked my customers "could you please just install Chrome? It's free, fast, and it works." Hopefully the competitive pressure will mean that future IE users will no longer have to be second class netizens.

Re:I thought this was already refuted?

[gx5000]

Agreed, but we do have to set some standards whether we're setting up an Gov Intranet or website for the general public, there's the fun now right ?

Re:I thought this was already refuted?

[gx5000]

Agreed, except I'm still on Opera because of the inherent problems with Chrome/Google cabal.... I just don't have the time to support all my friends and family anymore, and as my profile indicates, getting to old to...

Google has this habit

[PartOfElite]

It's not only Chrome - they try to inflate Google+ user count also, by counting every single Google service - including search engine and YouTube - as part of Google+. Then they boast user counts of like 100 million while the users have been nowhere near Google+ itself and it's perfectly clear there's not that kind of users. It's part of their marketing.

Re:Google has this habit

[spacepimp]

That is part of the misunderstanding people have about Google+. Google plus isn't a Facebook competitor. The way Google has been spinning it is that it is the integration of all of Googles services into a more central account base. Youtube, maps, gmail, google+ accounts, gchat, google music, have been consolidated. they are all part of Google+. People want it to be a street fight between Facebook and G+, so they see it for what they want it to be. You can argue that Google muddies the water by doing this, but to not streamline these services is counter intuitive, and difficult to manage.
Before Steve Jobs died he met with Larry Page and offered advice. Cutting the cruft and tying their products into a cohesive ecosystem are likely the advice he offered.

Re:Google has this habit

[mcgrew]

And it's not only Google, MS does similar things. Take their search engine. I don't know how many times I've been sent to Bing when clicking on a link not remotely related to Bing. Does anybody actually use Bing on purpose?

Every company is sleazy, including Google. Some are sleazier than others, of course (IMO the sleaziest tech company is Sony).

Re:Google has this habit

[PartOfElite]

In that case it balances out because Google does exactly the same. If you search on their other services, then many of them send you to the search engine. Google also offers sites their own "custom search" things and widgets that can be used to search that specific site (which just sends the user to Google with site: parameter). Google actually does this far more than Bing.

Re:Google has this habit

[Xest]

That's nothing, Facebook has this habit of paying people to troll Google on Slashdot!

Re:Google has this habit

[Shadowmist]

That is part of the misunderstanding people have about Google+. Google plus isn't a Facebook competitor.

I remember that earlier versions of Picasa had options on sharing your photos with Facebook. Those options got yanked not that long before Google Plus was launched. So I don't think the idea of competing with Facebook is that far from the truth.

Re:Google has this habit

[Pieroxy]

And... Steve Jobs was never wrong?

Re:Google has this habit

[spacepimp]

That is part of the misunderstanding people have about Google+. Google plus isn't a Facebook competitor.

I remember that earlier versions of Picasa had options on sharing your photos with Facebook. Those options got yanked not that long before Google Plus was launched. So I don't think the idea of competing with Facebook is that far from the truth.

I guess the wording could have been more precise on my part. let me restate it: Google+ isn't just a social network. It is the comprehensive unification of Google services into a more tightly knit ecosystem. Does it compete with Facebook? Yes, but in thinking of Google+ strictly a social network to compete with Facebook is missing the bigger picture. Maybe they will become more alike in the future as Facebook broadens it's scope.

Re:Google has this habit

[Fastolfe]

What makes you think Google had anything whatsoever to do with this?

Re:Google has this habit

[spacepimp]

I didn't say or suggest a thing about right or wrong. The Steve Jobs reference was suggesting part of the impetus to consolidate and unify the ecosystem, and the focus on fewer more refined products. Indirectly it was meant to suggest that the move to Google+ had origins outside of a square off with Facebook.

Re:Google has this habit

[Dishevel]

Before Steve Jobs died he met with Larry Page and offered advice.

I think that the advice given was "Fuck off and Die. I will destroy your asses from the grave!"
At least that is in line with everything we had heard him say about Google before.
Steve Jobs turned into a self entitled little fucking brat. Sad really. He started out as an awesome dude.
Then he got full of himself and decided he never needed a lic plate cause he was special. That he could park in handicapped spaces because "I am Steve Fucking Jobs".
I do not like ego driven assholes ever really. But Steve started so high in my opinion and went and got so low that I have a special place of hatred for him.

Re:Google has this habit

[sarysa]

Frankly I don't blame them. The only way anyone is to crush facebook is by creating the perception that people are switching over. I say the blame belongs with natural herd menality.

Re:Google has this habit

[spacepimp]

Before Steve Jobs died he met with Larry Page and offered advice.

I think that the advice given was "Fuck off and Die. I will destroy your asses from the grave!"
At least that is in line with everything we had heard him say about Google before.
Steve Jobs turned into a self entitled little fucking brat. Sad really. He started out as an awesome dude.
Then he got full of himself and decided he never needed a lic plate cause he was special. That he could park in handicapped spaces because "I am Steve Fucking Jobs".
I do not like ego driven assholes ever really. But Steve started so high in my opinion and went and got so low that I have a special place of hatred for him.

I think Steve was always arrogant, started out that way, and died that way. From what I've read he denied the request initially but was reminded how many people offered him advice when he was starting and then he accepted. When faced with death people often look at the world a bit differently. My father was an angry man most of his life. When the reality of mortality became unavoidable it broke down a lot of barriers in his thoughts, and relationships with friends and family.

If anything in my comment was seen as a praise or support of Steve Jobs, it is your own hate clouding the interpretation. There is plenty to hate about Steve Jobs, Microsoft, Facebook, Sony, Google and especially Oracle.

Re:Google has this habit

[arkane1234]

Doing that is about like a democrat joining the republican party because the democrats are too political...

Re:Google has this habit

[alnjmshntr]

I use Bing, simply because I also use Gmail and I don't want Google to be able to build up a huge profile on my web activity.

Really Bing is not at all bad, though the one thing I miss is the Google calculator.

Re:Google has this habit

[twocows]

I use Bing Rewards, actually. I don't care much what search engine I use as long as I get decent results and having Microsoft pay for my gaming habit doesn't bother me one bit.

Re:Google has this habit

[JustSomeProgrammer]

2 characters, but who bothers to count past 5 anyway?

Re:Google has this habit

[TangoMargarine]

because "I am Steve Fucking Jobs".

Well, at least he wasn't "Fucking Steve Jobs."

Re:Google has this habit

[ancientt]

(sipping KoolAid)

iGoogle is fine as a portal and I like having my Facebook friends statuses in it via RSS, recent email plus my daily news interests, but Google+ has status information, tracking of associates, share controls, auto-backup of my pictures and sites with full blown CMS.

Re:Google has this habit

[dbcad7]

Maybe I have become paranoid, but unless a link is obvious I am usually checking the bottom of the browser to see where a link is sending me before I click it. I don't think there are that many redirects to Bing or Google.., but sure there could be many to Bing or Google search results, but those as well as direct links are easy to spot.. Broken links should lead to Web page not found errors, although I know some configurations can lead to your default search engine.. If Bing is your default, then there you go.

Re:Google has this habit

[St.Creed]

It's not herd mentality, it's the simple fact that a social network is only as useful as the number of people that are on it. If telephone networks could only call within the network, everyone would join the largest one they could find.

Re:Google has this habit

[sarysa]

It's not herd mentality

it's the simple fact that a social network is only as useful as the number of people that are on it.

aka, herd mentality. Not picking something because it's the best in their eyes, but because more people are using it.

Re:Google has this habit

[ThatsMyNick]

No they do not. When you sign up for a google account, there is a check box, by which you can choose to join or not join Google+ (it is set to join be default, but it is optional). For older accounts, you have explicitly join Google+ to be part of Google+. And they do not count you as a Google+, unless you really have a Google+ account (well they are not idiots, and would not want to be sued for false advertising).

Re:Google has this habit

[mcgrew]

I had Google search boxes on my sites. The difference is that when you use one of these Google search boxes, you know you're using Google; the box is clearly marked with the famous G. With Bing, that's not the case. You're sent there by stealth.

Re:Google has this habit

[mcgrew]

I have a Gmail account and a G+ account, but I rarely check my Gmail and haven't been to G+ in months. I like Google's search because I get better results. If Bing gave better results than Google I'd switch like I switched to Google from Infoseek.

Re:Google has this habit

[Branciforte]

That had more to do with reciprocity. Facebook sucks in a lot of data and does not share back. Google used to let people export their information to Facebook (contacts, etc...) but Facebook would not to the same in return, so Google eventually got fed up with feeding the walled gardens.

Re:Google has this habit

[Branciforte]

Um, that's what Google+ is. Google+ is the integration of all the Google services into one unified whole. So, for instance, you can do a Google+ search and see results that include you email, docs, photos, etc...

There is also a sharing page that people seem to think of as Google+, because that is where a lot of the integrated services come together. And they media tends to portray the social page as being all of Google+, so them they can make wild stories about Google and Facebook competing with each other. And I am sorry that the media misled you, but the reality is that Google+ is the integration of all the Google services into a seamless whole.

Inside of Google, there is not much talk about Facebook, or competing with Facebook. Most people at Google could care less about Facebook, except for Facebook's astroturfing. Most of the recent focus at Google is on integrating everything and making all the service interact with each other. And the number one place Google likes to show off that integration is on what you call the Google+ page. But when Google publishes numbers for Google+, they are talking about the number of accounts that use the integrated services.

Re:Google has this habit

[Shadowmist]

That had more to do with reciprocity. Facebook sucks in a lot of data and does not share back. Google used to let people export their information to Facebook (contacts, etc...) but Facebook would not to the same in return, so Google eventually got fed up with feeding the walled gardens.

Save that it wasn't Google's stuff that was being shared it was mine. I was making the choice to upload my photo albums to my Facebook page. Not Google giving it away. What they were doing was blocking MY choice, not their own. Google is more like the petulant child who's walked out of Facebook's party and now setting up their own trying to inhibit his buddies from going to their party of choice.

Gotta love the commentary last sentence

[Anonymous Coward]

Regardless of how the market share was gained, duplicate counting inflates numbers. Get over it, Timmy.

Kinda funny how editors don't give a shit about editing, but when they want to put in some of their own editorial commenting that have no problem with it.

Re:Gotta love the commentary last sentence

[squiggleslash]

I kinda felt bad for the editor on that one. It felt like some school yard debate that ends with the loser saying "Oh yeah? Oh yeah? Yeah? Well, uh, at least I don't have a gay haircut".

Come to think of it, I'm a little concerned our next President might be someone who once did actually debate like that.

Re:Gotta love the commentary last sentence

[nedlohs]

How is eating dogs morally different from eating cows or pig or salmon?

Heck, dogs are the go to meat/bones supply for my fortresses of dwarfs.

Re:Gotta love the commentary last sentence

[mstefanro]

Dogs are more cute.

Re:Gotta love the commentary last sentence

[oreaq]

Dogs have personality, personality goes a long way.

Re:Gotta love the commentary last sentence

[ColdWetDog]

Dogs have personality, personality goes a long way.

If that were a criteria, cannibalism would be rampant.

Re:Gotta love the commentary last sentence

[nedlohs]

So do pigs, geese, and crickets. Yet eating them is fine (well unless you are Jewish/etc).

Re:Gotta love the commentary last sentence

[John Sugs]

Dogs have personality, personality goes a long way.

Other animals most definitely have personality too. I have a gecko family living on my balcony. They are all quite different, and actually remarkably wise too. You know the sound that gecko does (sounds like their name exactly, 'gecko'). It's used to call other geckos and their family members and friends.

I also have a mutual, good relationship with them. They can live in peace on my balcony and roam inside my apartment from the little space under the balcony door as they want, as long as they don't hang around here all the time. In turn, they take care of getting rid of any possible insects and flies. They seem to respect that quite well. Apart from the baby gecko who gets himself stuck somewhere and then calls his parents. Usually I just help him out and then he roams back to the balcony.

Re:Gotta love the commentary last sentence

[oreaq]

http://www.imdb.com/title/tt0110912/quotes?qt=qt0447129

Re:Gotta love the commentary last sentence

[oreaq]

I don't eat gecko either.

but coming pre-loaded...

[ganjadude]

is not at all artificially inflating the numbers.

or what about MS specific webapps such as their CRM system? I mean I could see if opera were the company that was making the complaint.

Re:but coming pre-loaded...

[PartOfElite]

What are you talking about? It's not about coming pre-loaded and bundled with every piece of shareware and freeware software, it's about preloading links and web pages as user browses the internet..

Re:but coming pre-loaded...

[PartOfElite]

It doesn't necessarily make it faster. It eats up lots of bandwidth. I'm on a poor connection and doing stuff like that slows down the whole internet.

It's also dubious that Google boasts their market share with these inflated numbers...

And what if one of the search results contain child porn or something else illegal? That's a nice way to get you to be suspect and your home raided.

Re:but coming pre-loaded...

[ganjadude]

I was not very clear, I was actually talking about IE coming preloaded on windows causing MS marketshare to be artificially high. My apologies, its early

Re:but coming pre-loaded...

[Dan East]

The method in which a user ends up with a browser - by default or by choice, etc - is a whole different topic. What is important for web developers are accurate statistics. I agree with MS on this one, because it sounds like the stats were quite skewed by page preloading, etc. How people ended up with IE doesn't change who is actually using what. I'm trying to figure out why Firefox and Chrome usage is so low on iPad devices - it's quite an anomaly - but again, that's a whole different topic.

(to save those who don't grasp subtle sarcastic humor, my comment about iPad browsers is totally tongue-in-cheek)

Re:but coming pre-loaded...

[Sepodati]

>> I'm on a poor connection and doing stuff like that slows down the whole internet.

Yes... I noticed the entire internet slow down when you searched earlier. Please stop.

>> It's also dubious that Google boasts their market share with these inflated numbers...

So you think it's doubtful or questionable that Google does this? So do I.

>> And what if one of the search results contain ...

OMG THINK OF THE CHILDREN!!

At least you earned your shill paycheck today.

Re:but coming pre-loaded...

[im_thatoneguy]

If you really care about metrics though you shouldn't be looking at StatsCounter's global stats anyway. Most likely your demographic is going to be unique and particular to your subject matter.

Lots of corporate viewers? Expect lots of IE.
Lots of gadgeteers? Expect lots of FF and Chrome.
Lots of Linux geeks? Expect lots of FF and Opera.

Etc...

You really need stats for your specific genre of website to best target your users. I would imagine that Slashdot's view stats are skewed pretty dramatically off of global.

Re:but coming pre-loaded...

[Patch86]

Coming preloaded so that people use it isn't artificially inflating the statistics- it's gaining real honest to god users (through sleazy and unfair methods). Storming into people's houses and forcing them to use your software at gunpoint would be sleazy and unfair, but there's no denying that they are now technically your users...

Re:but coming pre-loaded...

[Johann Lau]

Yes... I noticed the entire internet slow down when you searched earlier. Please stop.

You wouldn't know the difference either way, so that attempt at sarcasm is kinda backfiring.

So you think it's doubtful or questionable that Google does this? So do I.

I know that's what you think, but what do I think?

OMG THINK OF THE CHILDREN!!

At least you earned your shill paycheck today.

It's rather "OMG think of Joe Sixpack" but hey... while you're being dumb you might as well go for 3 out of 3.

Re:but coming pre-loaded...

[Johann Lau]

Statistics are useless anyway, not in the least because of all those bots that pretend to be browsers. Accurate statistics would be nice, but since they're not available, a web developer should spend more attention to what they are actually doing, IMHO... and who, generally and exluding situations that involve terminal greed or idiocy, cares about how many people use old version of IE? I don't, that's like looking at a turd before you flush. I mean sure, have a glance, but don't keep looking and looking and looking... remember what Nietzsche said ^^

This makes sense

[The Dancing Panda]

It seems like kind of a quick jump otherwise.

It really does not matter...

[hackula]

...Lynx rules all the browsers anyway.

Re:It really does not matter...

[crazyjj]

Am I the only one still using Cello?

Hello...anyone?

Re:It really does not matter...

[dkleinsc]

Lynx? Luxury. We had to do everything with telnet.

Re:It really does not matter...

[Nimey]

Probably. I couldn't even get it to run on any of my "vintage" virtual machines, let alone render any websites.

Re:It really does not matter...

[kno3]

And you try and tell the young people of today that ..... they won't believe you.

Re:It really does not matter...

[buchner.johannes]

elinks has mouse and javascript support. Just saying.

Re:It really does not matter...

[cant_get_a_good_nick]

At one point at school I was using Charlotte a text mode webbrowser on a mainframe. I also maintained a price list on our webserver, and I had to test on Charlotte to make sure it was readable there.

Surprisingly it seems to still exist and updated shockingly recently.

Re:It really does not matter...

[brianerst]

Lynx? Doesn't it use that weird WWW thingy?

I'll have to look it up on Gopher and see what all the fuss is about...

Wait a second

[Yebyen]

As usual, the summary makes no sense at all.

So, Google Chrome users who search on Google are counted as users, but they should not be counted?

Or, they are being counted twice? Or are they being counted for the number of tabs they have open?

What's an "invisible tab?" I don't want to read the article, but I don't understand how it inflates the actual number of chrome users. If the summary indicates what the article actually says, then there's no reason to discount these users, as they're not "actually not running Chrome"

Hanging chads!!!!

Re:Wait a second

[Anonymous Coward]

Chrome automatically loads some of the links on the page you are reading in the background, so that when you click on one of those links, it already has the page mostly ready. So when the user reads one page, "the web" sees several pages being loaded.

Slashdot 10 years later, what has changed. Microsoft still the evil empire, Google still the darling startup, and nobody can be bothered to read the article when it's about evil M$.

Re:Wait a second

[MyLongNickName]

What's an "invisible tab?" I don't want to read the article, but I don't understand how it inflates the actual number of chrome users

I think you said it all right there...

Re:Wait a second

[Yebyen]

Did you look at the article? Geolocation weighting? It's bloody five pages.

I don't come to Slashdot for the articles :)

Re:Wait a second

[PartOfElite]

Neither StatCounter or Net Applications have access to all the sites on the internet. How would they? Instead, they track data using their own analytics services and such that webmasters can use. By preloading every page on search results and when moving around the internet, Chrome greatly increases their changes of being picked up those statistics services.

On top of that it also directly inflates the usage amount as now there's 10x more hits than with other browsers.

Re:Wait a second

[Anonymous Coward]

I don't come to Slashdot for the articles :)

Then you'll have to settle for snarky responses. Glad to oblige.

Re:Wait a second

[Yebyen]

Don't those analytic services load an IFrame or some trick enough to get at their own statcounter cookies, and uniquely identify the user (seat)?

Yes, it would increase the chance that you get picked up and noticed as a user, but I don't see this being a problem for IE, since you only have to be noticed once to get counted. This "invisible tabs" isn't something I see throwing the count into 10x territory since that user would still be uniquely identified by the tracking cookies. So I visit 10x as many sites (silently, in the background) as a Chrome user, generating more traffic.

I'm not spawning guest VMs that have additional instances of Chrome, and keep their own separate cookie stores. (OK, so I am, but not because Chrome did it for me. This is not a feature of Chrome. IE users are probably equally likely to have extra IE machines.) Why does any of this make me any more difficult to identify and count by IP/cookies, or skew the numbers in favor of Chrome?

Re:Wait a second

[ProfessionalCookie]

But if you're on Chrome, you read it, even if you didn't :)

(but StatCounter didn't count you)

As opposed to what? Naturally boosted on the web?

[BMOC]

There's something "natural" on a completely artificial construct?

Wrong conclusion

[Old97]

In any case isn't what is being counted is page loads, not users? IE has more users because most computer users have IE. Mozilla and Chrome users may be more savvy and may actually use the internet more than IE users. It makes sense that if you care that much about the internet you probably have strong opinions about the browser you are using. If your company is trying to reach users you may want to know the percentage of users each browser accounts for. If your company is trying to reach the more engaged users you might be interested in page loads. This is similar to the android/iphone marketshare nonsense. Iphone users access the internet from their phones more than android users. More people own android devices than iphone devices. Everyone claims victory.

Re:Wrong conclusion

[PartOfElite]

Since neither StatCounter or Net Applications have access to every site on the internet, Chrome's preloading greatly increases their change of being picked up by either service. This inflates the numbers.

Re:Wrong conclusion

[DragonWriter]

In any case isn't what is being counted is page loads, not users?

Many firms measure page views, Net Applications (which shows IE in the lead) counts unique users.

Which is more important depends on what you plan on doing with the numbers (abstract "my browser is bigger than yours comparisons" not tied to any actual useful purpose can use either equally.)

Canadian stats

[GabboFlabbo]

Stats from a website which has mostly Canadian viewers:

Unique Users for the past 30 days
1.IE         66,554    42.21%
2.Safari     37,213    23.60%
3.Firefox    20,703    13.13%
4.Chrome     14,552    9.23%
5.Android    3,736    2.37%

*source: google analytics

Re:Canadian stats

[gislifb]

microsoft-shills.ca ?

I kind of agree

[Anonymous Coward]

I can't think of anyone's computer Ive seen recently that has been using Chrome, it's all Firefox or IE and one computer had Safari on Windows (Dont know why)

On The Other Hand, Could It Be...

[EXTomar]

Could it be that Chrome is on every Android platform and Android is on a lot of things? Many more pieces of hardware than Windows Mobile. Although I am a little dubious of the claim that "Chrome is #1" the growth makes a lot of sense where it has nothing to do with "hidden tabs" but that the installbase has exploded.

Re:On The Other Hand, Could It Be...

[Loosifur]

I've used a Droid, an Eris, and now have an Incredible. I use the Chrome sync feature, and I seem to recall that carrying over from the Droid up to the Incredible. Are you sure the default browser isn't Chrome, maybe just without the branding?

Re:On The Other Hand, Could It Be...

[ProbablyJoe]

It's Webkit based, and that's about where the similarities end (so is Safari and a lot of minor browsers). You could say the same about Chrome for Android too really, there's nothing particularly "Chrome" about it besides the branding. Not sure what sync feature you're referring to, but I think it's probably more of a Google account thing rather than a Chrome browser thing, there were definitely things like that before Chrome was released for Android.

The User Agent doesn't specify it as Chrome anyway, and I believe that they count mobile browsers seperately. Even if it is included, Android browsing only accounts for about 2% of the total (http://en.wikipedia.org/wiki/Usage_share_of_web_browsers#StatCounter_.28July_2008_to_present.29), so it's not really significant enough to sway things.

Re:On The Other Hand, Could It Be...

[Sancho]

I got a Droid about 3 months after they came out. They definitely weren't using Chrome. It was one of my gripes. Chrome for Android was released in Beta 4-5 months ago. Before that, you just had Browser. Both can coexist on a phone.

The fourth FAQ here indicates that they are different: https://developers.google.com/chrome/mobile/docs/faq

Re:On The Other Hand, Could It Be...

[BZ]

1) The Android browser is not Chrome (different UA string, different JS engine, different WebKit version, etc).

2) Total smartphone internet usage is much much smaller than desktop usage, so numbers that measure usage as opposed to installs are still pretty desktop-dominated.

Re:On The Other Hand, Could It Be...

[0100010001010011]

You can install Chrome in ICS. I have it running on my Asus Transformer right now. It's a bit different than the bundled browser but 90% the same.

Re:On The Other Hand, Could It Be...

[BZ]

Sure. You can install the Chrome beta. But the point is that it's not installed by default, install numbers are low, and ICS deployment is _also_ low.

All of which is to say that Chrome on Android is not a significant part of Chrome market share.

Re:On The Other Hand, Could It Be...

[bill_mcgonigle]

Could it be that Chrome is on every Android platform and Android is on a lot of things?

Chrome is just a shakey beta. IIRC it requires ICS and OpenGL to even work.

Re:Dumb

[Millennium]

One could argue that the bundling was the "behind-the-scenes shenanigans to inflate their numbers," particularly given common browser bundling practice at the time (also known as not doing it). That argument would be much weaker in today's environment, where everyone bundles a browser, but Microsoft's decision was not made in that environment.

Wikimedia stats agree with StatCounter

[David Gerard]

The Wikimedia browser stats pretty much match the StatCounter ones: 25.36% IE, 24.99% Chrome.

Note that Wikimedia is (a) a top-10 site with a broad general international readership (b) a charity with no direct interest in the question of "which browser wins?" but only in knowing the actual answers, so as to serve the readers.

Re:Wikimedia stats agree with StatCounter

[Anonymous Coward]

The Wikimedia browser stats pretty much match the StatCounter ones: 25.36% IE, 24.99% Chrome.

Note that Wikimedia is (a) a top-10 site with a broad general international readership (b) a charity with no direct interest in the question of "which browser wins?" but only in knowing the actual answers, so as to serve the readers.

Nevertheless, their stats could still be inflated artificially by browsers doing pre-rendering, which is where the discrepancy is claimed to come from.

Re:Wikimedia stats agree with StatCounter

[PartOfElite]

Wikipedia comes up in almost every search query on Google. Chrome heavily inflates their numbers too.

Re:Wikimedia stats agree with StatCounter

[David Gerard]

Quite possibly, yes. Those are raw numbers from the Squid caches (a sample of 1 in 1000 hits).

Re:Wikimedia stats agree with StatCounter

[Kalriath]

Because Chrome (and Firefox with pre-fetching enabled) will pre-download the top x results - 10 I think. The search engine used is irrelevant, the browser is the issue at hand.

Re:As opposed to what? Naturally boosted on the we

[Millennium]

If we're going to be pedantic, would you prefer "emergent vs. engineered" to "organic vs. artificial"?

NETSCAPE

[rullywowr]

I am typing this on netscape at the moment and I totally agree with th ^^^^NO CARRIER

Re:Prerendering

[DocSavage64109]

I was thinking the same thing. The stats indicate that Chrome is using more bandwidth and resources than IE despite having 1/3 the install base. Seems inefficient.

Just adblock lowlives like StatCounter

[KiloByte]

You can't use a browser without adblock these days and retain sanity. And unless you decide to throw away your privacy, you'll block trackers like Google Analytics or StatCounter.

So join me on the mission: drive apparent Firefox usage stats to 0.

Geo Weighted

[medv4380]

What is country level weighting, and why do you do it?
The Net Market Share data is weighted by country. We compare our traffic to the CIA Internet Traffic by Country table, and weight our data accordingly. For example, if our global data shows that Brazil represents 2% of our traffic, and the CIA table shows Brazil to represent 4% of global Internet traffic, we will count each unique visitor from Brazil twice. This is done to balance out our global data. All regions have differing markets, and if our traffic were concentrated in one or more regions, our global data would be inappropriately affected by those regions. Country level weighting removes any bias by region.

So I'm to trust numbers that I know have a flawed methodology?
Why not these then
http://www.w3schools.com/browsers/browsers_stats.asp
Oh, but weight, stat counter started removing chrome over counts.

Further to a significant number of user requests, we are now adjusting our browser stats to remove the effect of prerendering in Google Chrome. From 1 May 2012, prerendered pages (which are not actually viewed) are not included in our stats.

Re:Geo Weighted

[edxwelch]

The CIA data is for number of internet users, not internet traffic, where internet user is some one who uses the internet at least once a week. And that's the flaw. the China stats are weighed upwards, because they have much more internet users, but the average Chinese user may be only checking his e-mail the odd time in an internet cafe, whereas internet users in western countries are constantly online.

Other sources agree with Statcounter

[INowRegretThesePosts]

See http://en.wikipedia.org/wiki/Web_browser_market_share#Summary_table

In the data for April, only Net Applications put MSIE significant ahead of Google Chrome. The other 3 sources, on average, give *lower* usage of MSIE than Stat Counter.

Re:Other sources agree with Statcounter

[Kalriath]

And what did you do, tech support? Yes, I'm sure you are fully aware of the contractual arrangements between Microsoft and your company's executives, and how much money moved in which direction.

More likely is that Microsoft just reminded your company that if they use more Windows, they'll only end up paying like $0.30 a month each license.

Aggressive promotion

[slasho81]

Eventually Chrome will rule the market. Google promotes it aggressively from their homepage. Under today's musical doodle was this text: "Upgrade to a modern browser and see what this doodle can really do." I'm on Firefox 12, by the way.

Re:Aggressive promotion

[Kalriath]

That's OK, I'm using Chromium and I don't even get a doodle on their homepage.

Re:As opposed to what? Naturally boosted on the we

[BMOC]

Actually yes, and I'm not ashamed of my preference.

The statistics are done by VISIBLE tabs???

[gradinaruvasile]

I was under the impression that counting browser share is done by taking a sites visitors and identifying the ACTUAL number of users of each browser. But then i read that page hits are counted and hidden tabs are skewing results. Ok , so i use browser X and i have tons of tabs open, HIDDEN OR NOT, does that mean that i somewhat contribute to artificial inflation of my browsers share? Because as i understand it browser share should indicate the usage of a browser by actual people and NOT the amount of VISIBLE tabs opened by them. By this metric, people like my wife are essentially skewing the results (she has the habit of opening 50 or more tabs)? Additionally netstat uses weighting that takes this whole counting thing into a more "controllable" area - just downgrade country x, upgrade country y and voila you have a few percent +/- (well not that simple, but anyway). I am no expert in statistics but i know that when different stuff are factored in, the greater the chance of adjusting the results. Personally i see that fewer people use IE - mostly company execs/workers who are under the impression that it is actually good because its made by Microsoft or because the computer came with it and they dont "have the time" to "learn to use" something safer and faster - most users actually use Firefox or Chrome.

Does all software lie now?

[fredrated]

In the 'old' days when I worked with software, I knew what I had. Punch cards, core dumps, later DOS, wysiwyg.
Now everything lies. Windows hides stuff, Linux hides crap, browsers hide stuff. How the hell is one supposed to know what is going on, particularly when trying to troubleshoot/debug stuff when even the software lies?

By the way, I don't know what software slashdot uses, but trying to write this post was annoying, I kept loosing the insertion point and couldn't navigate the text!

Re:Does all software lie now?

[mcgrew]

I kept loosing the insertion point and couldn't navigate the text!

Well, there's your problem right there. Rather than loosing it you should have tightened it.

MS IE dependent

[linux.py]

MS IE has to be installed for Windows to work. This is not the case for any other web browser. I suspect that if MS removed the dependencies IE would be removed from many more computers. Invisible tabs or not, Chrome is just a more pleasant and useful browser.

Re:MS IE dependent

[Kalriath]

IE does not need to be installed (and can be removed). Trident does need to be installed and cannot be removed. And if you complain about that, remember that OS X won't work without WebKit, and KDE won't work without KHTML.

Re:MS IE dependent

[linux.py]

All I know is when I uninstall IE Windows doesn't function properly anymore. I'm sure there is a way to remove it while keeping the necessary libraries, but that is beyond most users. My point is that Chrome is not the default browser for any standard OS and there are no implicit dependencies yet it still has a larger user base. OS X has Safari and most linux distros come with Firefox. Chrome is usually installed beside another default browser and it still gets used more often.

Re:MS IE dependent

[Kalriath]

Windows 7 and higher are the ones where you can remove IE. And it does leave Trident behind. That said, I know Linux will happily let you remove the default browser, but does OS X?

Preloading and employer filters

[Mr Z]

A tangentially related question: Has anyone gotten in trouble with violating their employer's Acceptable Use Policy due to browser preloading / precaching? Often, in search results or even certain news sites there are outbound links to places I'd never visit from work. But if Chrome (or even Firefox) is clicking those links behind my back, my IP address is in a corporate log somewhere as having "visited" that site, isn't it?

How are these preload/precache "hits" distinguished from normal hits? Obviously, if some of the sites are filtering these out, there's some way to tell them apart. At the same time, if the "hits" were noticeably different, there's always the chance the webserver would serve up different pages based on this difference.

Re:Preloading and employer filters

[Mr Z]

It also raises the ancillary question: If my browser does precache/prerender a page, how does the website detect when I do actually visit it?

Re:Preloading and employer filters

[WhyDoubt]

For Chrome, there is a related API that both NetMarketShare and StatCounter are using to figure out what is really happening.
http://netmarketshare.com/prerendering.aspx
http://gs.statcounter.com/faq#prerendering

bonch has this habit

[AliasMarlowe]

That's nothing, Facebook has this habit of paying people to troll Google on Slashdot!

Possibly not in this case. The person who posted the story was bonch, who appears to post questionable stuff in favor of MS and against Google.

MS at it again

[Tom]

Imagine my surprise when I loaded up Skype on my iPhone today and noticed that at the top of your contact list, it now displays an ad banner - for Internet Explorer! With an "install now" notice.

Not only did I think someone at MS might be smart enough to realize that I can't install IE on my iPhone, but I thought this is the exact anti-competitive behaviour they had been found guilty for? You know, pushing the crapware IE with their near-monopoly in other areas?

Anyone got a Skype alternative? I knew it was time to dump it when MS acquired it. It was such a nice piece of software. :-(

So yeah, the IE market share is all organic. You know, as organic as plastic wrapped in shrinkwrap foil.

Re:MS at it again

[shutdown -p now]

Not only did I think someone at MS might be smart enough to realize that I can't install IE on my iPhone, but I thought this is the exact anti-competitive behaviour they had been found guilty for? You know, pushing the crapware IE with their near-monopoly in other areas?

I doubt Skype has a monopoly in its market.

Browser preloads an invisible tab?

[dgharmon]

> When a user enters a search in Chrome, the browser preloads an invisible tab not shown to the user, and these were being counted by StatCounter ..

What's the name of this ' invisible tab`?

It is a sorry state

[youknowwhat]

Give me any "decent" company that survives over 10 years that never "boosts" any statistics, never "sues" its competitor with stupid patents, never kill its competitors using its monopoly power......

Poppycock!

[lymond01]

I run it on 100% of my two machines!

Re:Poppycock!

[arkane1234]

Same here... and IE is only used for plugins from vendors when it's IE only.

Who cares

[wjousts]

Seriously? Use whatever browser works for you. Who gives a shit about this stupid pissing contest.

IE usage is Inflated (Windows Update)

[jameskojiro]

IE only runs for a lot of people when they run windows update.

Because no one has found a way to use Chrome or Firefox to run windows update.... yet..

Re:IE usage is Inflated (Windows Update)

[kyrio]

Windows 7 doesn't use a browser for updates. Vista probably doesn't use a browser for updates, either. For many years now, there has been a site that gives you Windows updates with Firefox.

Yawn

[jacobsm]

And the world cares, why?

Microsoft browser usage artificially boosted

[Tough Love]

By bundling Internet Explorer with every copy of window it ships to its indentured OEMs, Microsoft artificially inflates its browser share.

We know the truth here

[3seas]

Pot calling the kettle black...

Re:I actually discovered Lynx is useful!

[hackula]

vim/irssi/lynx

This is my dev environment about 70% of the day now. C# the other 30 pretty much rules that set up out the rest of the time (I am no masochist). A strict command line environment can be great, even on Windows, when your platform does not have great tools anyway (wacky low level C-variant in my case).

I don't trust anyone

[Karmashock]

Everyone is biased as hell.

Sure, MS is trying to make IE look more popular somehow even though it's the crap browser. But then look at all the people that don't really care what the data is they just want to dog pile on Microsoft?

I hate these issues. People either need to be objective, detached, and open minded or they need to stay the f' out of statistics. Just go bang rocks together somewhere else with the other barbarians.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Yes, it is.

[kyrio]

Whereas the saturation of MSIE is totally organic, right?

Other than the possibility that some small amount of spam bots might be hitting some sites and identifying as IE, yes, it is organic.

Re:MS sucks

[kyrio]

IE(9) is currently better than Firefox. Opera is still the best out of all of them. Chrome might be somewhere in the middle.

Re:I thought we were talking about users??

[kyrio]

I'm not sure what you're saying here. I code my sites to be 100% XHTML/CSS3 compliant. IE(9) renders the pages just as well as any other browser.

Re:I thought we were talking about users??

[kyrio]

Aww shit, above was a reply to: http://tech.slashdot.org/comments.pl?sid=2869437&cid=40088329

not all visits are through search anyway

[dr_blurb]

When a user enters a search in Chrome, the browser preloads an invisible tab not shown to the user, and these were being counted by StatCounter. Net Applications, another usage tracking group, ignores these invisible tabs and reports IE at 54%, Firefox at 20.20%, and Chrome at 18.85%."

Is this a slashvertisement for Net Applications? 54% for IE?? Did they grab their data from 2009?

Also, not all traffic is search traffic. The stats for the last two months at http://www.calcudoku.org/ (which has < 35% search traffic):

1. Chrome: 31%
2. Firefox: 25%
3. IE: 24%
4. Safari: 17%
5. Opera: 1%
6. Android: 1%