Google Chrome's breakneck pace: innovation or version inflation?

Google Chrome's breakneck pace: innovation or version inflation?

Summary: Delivering two new packages every calendar quarter and assigning a major version number to each one makes it appear that Google is innovating at lightning speed while its rivals plod along. But what's really inside those releases? Is Google just practicing version number inflation?

SHARE:

According to the About dialog box, my installation of Google Chrome just rolled over to version 16. When the year began, it was on version 8.

That's eight new versions in one year, an average of one new "major" release every six weeks, not counting the two or three bug fixes that Google's updater delivers in between the big milestones.

Delivering two new packages every calendar quarter and assigning a major version number to each one makes it appear that Google is innovating at lightning speed while its rivals plod along. Indeed, that numbering scheme was enough to convince Mozilla to change its release schedule. Enterprise customers weren't very happy when Firefox adopted the same new-version-every-six-weeks model.

But one thing has been bugging me all year. Despite the rapidly incrementing major version numbers, Chrome doesn't really seem to have changed all that much this year. Are those formal releases really all that different? Or is Google's "innovation" really just a new form of inflation, targeting version numbers?

This morning I shared a wisecrack on Twitter that turned out to be more truthful than I realized:

Indeed, the closer I look at what's really in each new Chrome release, the more I'm convinced that most of these releases are pretty minor in the grand scheme of things. The move from version 15 to 16 really does look more like a minor upgrade from 1.5 to 1.6.

That's certainly true if you look at support for HTML5 standards. Given the constant drumbeat over the importance of HTML5 to the future of the web, one would think it's a key focus for Chrome developers. And yet all that incrementing has barely made a difference in HTML5 compatibility this year, as measured by independent test sites.

Exhibit A: The HTML5 Test site runs browsers through a battery of tests and then assigns a score to each one. A perfect score would be 450. Here's how the last six versions of Google Chrome have fared:

  • 11.0.696:     338
  • 12.0.742:     339
  • 13.0.782:     341
  • 14.0.835:     340
  • 15.0.874:     342
  • 16.0.891:     343

Those are downright microscopic changes over a period of roughly eight months, with only one or two tiny features at a time being changed. And the gap between HTML5 today and the future ideal is still massive.

Likewise, the Browserscope summary scores, which track functionality for web developers, include three versions of Chrome, 15 through 17, with identical overall scores of 87/100 for each one. On the Browserscope HTML5 2.2 test, compatibility scores actually decreased between versions 15 and 16, dropping from 332 to 317. That's the sort of thing you expect from a beta release, not from an upgrade that should be pushing standards support forward.

Looking beyond HTML5, some of the change logs for recent Chrome releases (as described in its Wikipedia page) are so thin that they make a mockery of the major version number strategy. In version 11.0.696, for example, only two additions appeared on the list: "HTML5 Speech Input API" is useful to developers but arguably a yawner for end users. The second change? An "updated icon." I would expect much more from a release that goes to 11.

Image via Wikipedia

Some of the changes listed for this year's whole-number releases are not so much about the web as they are about Google's services. Version 13.0.782, for example, introduced Instant Pages, a feature that downloads and pre-renders the top result in a Google search—it's a variation on prefetching techniques that have been around for a long time, and only of benefit in Google searches. The other substantive change in v13 was a change to the print preview interface. Again, hardly the sort of thing that would traditionally warrant a new major version number.

And when my ZDNet colleague Steven J. Vaughan-Nichols, a longtime Google fan, reviewed Chrome version 16 today, he acknowledged: "No, there’s nothing new in capital letters in this release."  It's slower than it once was, and the flagship feature in this release, multiple user profiles, seems half-baked based on Steven's description:

[Y]ou can sync multiple users to one copy of Chrome. So, for example, you can have multiple people via their Gmail accounts, running their own Chrome settings. ...

So far, so good, but, there’s no security between logins. She can see all my settings and I can see hers. You may be OK with that, but I’m not and I can’t see it in a work environment where people share PCs.

His conclusion? "at this point I don’t see [this feature] as being that useful." In other words, perhaps it should have been released to developers and early adopters, not to the general public.

Would the Web be worse off if Google decided to call each one of those a point release or even a beta? I don't think so.

Don't get me wrong. I think Google has in general done a great job with Chrome. It's fast, it has very few compatibility problems in my experience (although I regularly hear a few screams when a new version appears in the wild). And Google has done an admirable job with its auto-update process, which seems to work well—at least for consumers. But a predictable major release cycle of every six months, with optional updates every six weeks for developers and bleeding-edge users to experiment with new stuff, would be a far more sensible approach. That approach seems to work well for the Linux community.

Perhaps Google's motive with its fast-twitch release cycle is to unnerve its competitors. It seemed to do that with Mozilla, which is struggling to keep that pace. So far Microsoft hasn't taken the bait. It's still sticking with its annual release cycle, although it has moved to automatic updates in an effort to forcibly drag diehard IE6 and IE7 users into more modern browsers.

The bottom line? Next time someone talks about Google's rapid pace of innovation with Chrome, take that argument with a grain of salt.

Topics: Browser, Apps, Google, Software Development

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

48 comments
Log in or register to join the discussion
  • RE: Google Chrome's breakneck pace: innovation or version inflation?

    The difference between Chrome's fast paced speed, and Mozilla's is that Chrome's extensions *don't break* every release and cause frustration. Mozilla seems adept at breaking all of them each time a new beta comes out.
    The one and only, Cylon Centurion
    • RE: Google Chrome's breakneck pace: innovation or version inflation?

      @Cylon Centurion I haven't had any extensions, that I use, break since the move from version 2.x to 3.x.

      I like some of Chrome, but until NoScript or an equivalent is available, I won't be switching. I have tried some of the Chrome script blockers, but none offers the control and security that NoScript does - mainly due to the extensions API not supporting the required functionality that a NoScript add-on needs.

      Until NoScript becomes available for Chrome, I'll probably be sticking with Firefox as my main browser.
      wright_is
      • RE: Google Chrome's breakneck pace: innovation or version inflation?

        @wright_is
        There exists an equivalent, Not Script, for Chrome/Chromium. It is more complicated to use. Try Chromium, not Chrome, with Ghostery, Disconnect for some protection.

        I still prefer Firefox for more flexibility with tabs and other extensions.
        jorjitop
  • Is this an aberration or a trend Ed?

    Have software development methods (Agile?) at Microsoft kept pace with the rest of the world?
    Perhaps there is more than one way to look at this issue, yes?
    Dietrich T. Schmitz *Your
    • RE: Google Chrome's breakneck pace: innovation or version inflation?

      @Dietrich

      The internal organization of Microsoft isn't that simple. From what I've heard, they're split into many teams, and the individual teams tend to select how they do things. They're not really as monolithic as people make them out to be.
      CobraA1
      • Oh, you've heard? I think there is reason to believe you are blowing smoke

        @CobraA1
        Perhaps there is good reason for Google's and Mozilla's choice of versioning.
        Perhaps the pace at which they both code out-paces Microsoft, and that's is something worth mentioning since there isn't anything else Ed can find to critique these days.

        Putting the focus on Google takes the focus off Microsoft.

        Why aren't the programmers at Microsoft keeping up should be in the minds of intelligent readers.
        Dietrich T. Schmitz *Your
      • RE: Google Chrome's breakneck pace: innovation or version inflation?

        @CobraA1 It's not the internal organization what matters and what's a constraint here. Microsoft's internal structure is not the reason, but it's the result. It fit best what their users need and want. Because most users - including home and corporate ones - just don't give a damn about rolling version numbers and are not keen on being beta-testers of half-baken new features, so the developers can save on their own budget. They rather want mature, tested features, which they get on a regular basis from Microsoft. And they could never get *that* on a 6-weekly regular update basis.
        ff2
      • Look a little closer

        @Dietrich

        "Why aren't the programmers at Microsoft keeping up ?"

        You might try looking at the Platform Previews Microsoft releases. For a large portion of the past two years they have been releasing every six weeks. The difference is they are targeting developers, not equiring customers to upgrade before they're ready.
        Ed Bott
    • There are many ways to look at an issue, but in this case the issue is

      @Dietrich T. Schmitz * Your Linux Advocate
      that Google is in no sense releasing a new version of Chrome as the number system would suggest, instead releasing the same browser under multiple release numbers with minor fixes and upgrades.

      What this does is give those not well versed in software practices the illusion that Google is coding at a rapid pace, and that users are downlaoding and installing a major release, when in truth they are just downloading a minor bug fix or feature.

      That would be like Microsoft calling XP, Service pack 1 "Windows 7", SP2 as "Windows 8", and SP3 as "Windows 9"

      I think it is very telling about Google, that in truth they are coding at a slower pace then one should expect from a company that size, and feel the need to try and hide that fact.

      :|
      Tim Cook
      • RE: Google Chrome's breakneck pace: innovation or version inflation?

        @Mister Spock
        Logic would tell us that you failed to show any supporting facts to what you believe is true.
        daikon
      • Your logic is impeccable

        @Mister Spock
        +1
        :)
        William Farrel
      • Neither did the original poster

        @daikon

        Dietrich T. Schmitz - all he did was make an unsupported claim that

        [i]Perhaps the pace at which they both code out-paces Microsoft, and that's is something worth mentioning since there isn't anything else Ed can find to critique these days.[/i]

        I see nothing in anything that would support that, do you.

        Yet Spock backed up his opinion with some logic, so I'll go with his assesment as being the far more accurate one.
        William Farrel
    • RE: Google Chrome's breakneck pace: innovation or version inflation?

      @Dietrich T. Schmitz * Your Linux Advocate

      The development method has nothing to do with major version numbering...

      Just look at the Linux Kernel, it has just gone to 3.

      Chrome came from behind, while Microsoft were moving from version 7 to 8, Firefox to 4 and Opera to 10. That meant a lot of "version catching up", even if that is ridiculous.

      Changing an Icon doesn't make a major release! That is the point.

      Ed isn't knocking major version changes for major advances in functionality, like in IE and Firefox, up to version 4.

      But the current Google (and now Mozilla) method of using a new major version number for each release, whether it is major or not is getting silly.

      Mind you, so was the "constant beta" status of their software not long ago. "If we call it beta, we don't have to support it," is not a serious business model and made a mockery of companies that really did beta test products - customers assumed that beta test versions of products were actually full releases and complained loudly, when they found bugs... Well, duh! That's why it is a beta test, so you can find bugs and report them, before the final version comes out!
      wright_is
      • RE: Google Chrome's breakneck pace: innovation or version inflation?

        @wright_is This "major version numbering" stuff is what readers and the majority of ZDNet writers are just going to have to let go, like an urban legend debunked on Snopes. The number DOESN'T MEAN ANYTHING other than it's a new release. It's not like a telephone number or IP address that lets you look at the sub-numbers and infer some sort of useful information. It's JUST A NUMBER. Google didn't put out a press release and proclaim a major revision. They just released a new version. Mozilla just released a new version. When I wanted to know what was different in Mozilla I went to their website and read the changelog; I didn't contact a numerologist and pass them nothing but a version number.<br><br>"But the current Google (and now Mozilla) method of using a new major version number for each release, whether it is major or not is getting silly."<br><br>The only thing getting silly is that Ed, Adrian, and the rest (except Mary Jo Foley) continue to beat this drum with every. single. release. People explain to them that the numbering scheme contains no inherent information, they ignore this, and recycle the article again six weeks later.<br><br>We had this with OpenSUSE and people making a big deal about going from 11.4 to 12.1 rather than 12.0. The reality was that version numbers NEVER meant anything in OpenSUSE from its first release... other than 11.4, releases just went A.0, A.1, A.2, A.3, B.0 regardless of what went into them. Regardless, lots of users considered the .0 version a "major" version and would skip it because it's "too new" and wait for the "point release" (and encourage others to do the same). Meanwhile, press would consider the .1 release a "point/bugfix release" and give little or no coverage to versions ending in .1. Because of this craziness, the numbering system needed to be changed and .0 releases skipped from now on - and even that freaked people out.<br><br>PEOPLE - PLEASE LET THE NUMBERS GO. <br><br>This has been a public service announcement. Thank you. I now return you to complaining about things that don't mean anything and never did but you imagine they do anyway even after people explain this to you.
        jgm@...
      • Um, no

        @jgm<br><br>I am not "continuing to beat this drum with every release." I would be delighted if you would show me links to those articles you think I wrote. I certainly don't recall them.<br><br>On the contrary, it's our open source guy, Steven J. Vaughan-Nichols, who has *reviewed* every new recent version of Chrome. Even though, as you argue, he shouldn't be doing so.<br><br>And he is not alone. Here, try these searches:<br><br><a href="http://www.bing.com/search?q=%22Chrome+13%22+review" target="_blank" rel="nofollow">http://www.bing.com/search?q=%22Chrome+13%22+review</a><br><a href="http://www.bing.com/search?q=%22Chrome+14%22+review" target="_blank" rel="nofollow">http://www.bing.com/search?q=%22Chrome+14%22+review</a><br><a href="http://www.bing.com/search?q=%22Chrome+15%22+review" target="_blank" rel="nofollow">http://www.bing.com/search?q=%22Chrome+15%22+review</a><br><br>My point is about the PERCEPTION in the tech press about the extent of innovation in Chrome. Google, deliberately or not, feeds that perception with its insistence on using major version numbers.
        Ed Bott
  • RE: Google Chrome's breakneck pace: innovation or version inflation?

    "Google Chrome's breakneck pace: innovation or version inflation?"

    Easy: Version inflation. They're creating a new version for pretty much anything they do.

    Not that it matters, as they're intentionally not making version numbers important to Chrome. The version number is there for the developer, not for the average user. To the average user, it's just Chrome.

    "Exhibit A: The HTML5 Test site"

    I don't care much for that site. They pick and choose the standards they include, and include a bunch of non-standard stuff. It's not a real attempt to show any sort of comprehensive HTML 5 compatibility.
    CobraA1
    • Nonetheless

      @CobraA1

      Regardless of what you think about the HTML5 test site, they are pretty comprehensive, and the fact there's been no real HTML5 progress in 8 months or more despite all these new versions...
      Ed Bott
      • RE: Google Chrome's breakneck pace: innovation or version inflation?

        @Ed Bott No, it's not. Testing mostly form input types and a handful of various other features is hardly comprehensive. Forms is a tiny part of the specification, yet they're the largest part of the test suite.
        CobraA1
      • I understand your point

        @CobraA1

        I agree that HTML5test.com overemphasizes minor parts of the emerging HTML5 standards set. But they haven't left anything out. And the fact is that in eight months Google has done almost nothing to move the HTML5 standard forward in its browser. Agreed, the test doesn't fairly compare one browser against another. But it does allow one to see progress or lack thereof in a single browser.
        Ed Bott
    • RE: Google Chrome's breakneck pace: innovation or version inflation?

      @CobraA1
      Of course, HTML5 isn't a standard anyway.
      Fred Fredrickson