The Tuesday Grab Bag: Boot Time, Chrome v Firefox, and Personal Metrics

Share via Twitter Share via Facebook Share via Linkedin Share via Reddit

It’s pretty clear that in the battle between the seasons, summer lost. Badly. How else to explain the continuing foul weather and the spate of announcements that have kept us busy these past weeks? Normally, summer is both warmer and slower; this year’s version is anything but.

Which is why we’re scrambling a bit over here. In between conferences, the news is coming a mile a minute. And I’m not talking just about those minor pieces of news like the announcement of a new operating system. Just today VMware was reported to have poached Mark Lucovsky from Google, Sun had a rough quarter, Rackspace released a beta of its cloud API, and I’m sure a bunch of other things happened.

While we ponder those things and more here at RedMonk HQ, herewith are a few items that in typical grab bag fashion may deserve their own entries but aren’t going to get them.

Boot Time & the Smartphone

After I said my piece on Google’s Chrome OS launch to question, a couple of folks in conversation questioned my assertion that this is ultimately about speed; game changing speed. The primary objection was that with platforms like OS X supporting seamless suspend, the priority on boot time was misplaced.

I respectfully disagree. While I think that that objection is not incorrect, I think there’s something irrationally powerful in boot times that approach instant on, suspend or no suspend. People are used to computers – whatever the operating system – taking tens of seconds, if not minutes, to boot. Devices – netbook or otherwise – that can nearly eliminate that wait will seem disproportionately impressive, superficial though the gains might ultimately prove to be. In a world in which I keep my browser open more or less fulltime, is Chromium’s spectacular startup time really all that important? No.

It’s no less impressive for that, however.

I suspect, as well, that boot time will be at an increasing premium as smartphones begin to infringe on netbook/notebook/PC usage.

Chromium vs Firefox 3.5

Speaking of Chromium, it’s worth noting that I’ve been running the Linux builds of the project that serves as the foundation for the Chrome browser (Ubuntu users can grab it from the PPA here), and I have to say that I’m very impressed. It’s not usable for me as a primary browser yet, primarily due to the lack of basic functionality like Flash, but it is blindingly quick. Not just in the rendering – the benchmarks I’ve seen go back and forth on Tracemonkey vs V8 – but just the application itself. Firefox, which in its defense carries the much heavier load with all of the plugins I’ve installed, is definitely the tortoise to Chrome’s hare.

For all of the fanfare of Firefox’s 3.5 release – and kudos to them on a massively successful rollout – the browser, Tracemonkey notwithstanding, doesn’t feel that all that much faster. And it has a tendency to bog down by the end of the day, necessitating a restart which may or may not be practical depending on how many tabs I’m running.

Chrome’s speed hasn’t been enough, yet, to persuade me to switch, but it is highly differentiating.

Personal Metrics

Back in 2004, I wrote about what I then was calling “personal business intelligence,” an unwieldy term if ever there was one. Here’s how I described it:

The premise is absurdly simple: enterprise class tools essentially parse a wealth of corporate data and present it in some fashion – reporting, visually, or otherwise – to allow for identification of meaningful insights or observations. Their Personal counterparts would do the same, but with a different – and far less normalized, obviously – set of underlying data.

As consumers begin to accumulate metadata attached to their generated content – be it blogs, music, or photos, some interesting possibilities arrive for similar data mining. It’s intriguing, for example, to consider the possibility of a tool that can tell me – at a particular point in time – what I was thinking (via blogs), what I was reading (via bookmarks), where I was (via photos), who I was talking to (via email/IM), and what I was listening to (via music observation). Or to imagine a service such as I proposed in the link above, that provides me with more sophisticated intelligence on one area – like bookmarks – as the very interesting Extisp.icio.us does. In a sense, this is just extending the proactive rather than reactive use of metadata espoused by the Dashboard team, among others.

This is powered, of course, by not only the observable nature of many of our digital behaviors today – but by the ability to persist that information indefinitely.

Fast forward five years and that concept has largely become a reality, more so than I ever anticipated. As the folks from Wired cover, Nike has harnessed the power of personal metrics for thousands of runners worldwide, including me. More recently, Novell’s Nat Friedman talked about his and others efforts to “log” their lives, by which they essentially mean recording objective data about various aspects of their lives.

This phenomenon, which I think will dramatically accelerate in the years ahead if only because many of us will be carrying around devices that make the logging of metrics simpler, is interesting because of what it means about our ability to make data driven decisions about our lives. As Nat puts it:

We are bad at remembering our emotions and state of mind, and we forget daily events. We have theories about ourselves, but does the data match? I sleep too much; I never used to get sick this often; I’m incredibly hard-working; I’m a lot happier than I used to be. These are the things we tell ourselves, but without objective data, without a reliable memory of our past, how can we know if they are true?

But it is potentially more interesting because of the Hawthorne effect. As Wired’s Mark McClusky explains it, “the gist of the idea is that people change their behavior—often for the better—when they are being observed.”

More data about our lives means, then, that not only can we make better, data driven decisions about ourselves, but also that we might by sharing some of the metrics be incented to outperform our expectations.

Either way, one thing is clear: if the explosion of baseball data has taught us anything, it’s that quant skills are going to be at a premium as we go forward.


  1. Regarding boot-time, the reason it will appear to be important is because the vast majority of the population is used to Windows where suspend doesn’t work as well, crashes necessitate a reboot, and many of the automatic updates do too. With everyone used to having to reboot, a fast boot time will seem impressive. As long as Chrome OS doesn’t introduce stability issues not present in the Linux kernel today though, the boot time may become less important. Of course, if the boot time is so quick, it may make the need for suspend go away.

  2. re: personal metrics, the Hawthorne effect is pretty much the whole trick behind personal unit testing.

  3. The latest Chromium nightlies now support extensions and plugins on Linux. Just drop a copy of the Flash plugin (e.g. /usr/lib/adobe-flashplugin/libflashplayer.so) into /usr/lib/chromium-browser/plugins:


  4. regarding “boot time and smartphones” I am in violet agreement and I’m not one prone to violence. I have bulked up my iPod Touch with all of my important “tools” and it’s ability to turn a web app into a quick launch “icon” is all the more relevant. I can get to email, feeds, AP news, weather, and control systems in my home office all from the iPod and nearly all of them are the same tools I once used on my PC. The iPod (sitting in for the smartphone class) is nearly “instant on” (takes a second to find my network) and always ready. Its faster than pulling out the laptop even if the laptop is in suspend mode.

    Regarding “personal metrics” – the concept of a “life recorder” with temporal markers and full search capability is a reality for those of us wrapped around the axle of the digital age. It may be our salvation as we lose the ability to remember the basics in a sea of excess details. However, “reflection” on that data is still requires a significant conscious effort and it’s an effort few (me included) tend to affect.

  5. declarative living and counting.

    on boot times- word. i saw natalie boot her macbook the other day and i basically refused to believe she wasn’t coming up from suspend. that machine boots ungodly fast. i had some speedies envy i can tell you.

  6. […] My colleague Stephen O’Grady wrote a seminal feature some time ago – Speed is a Feature. I often return to this this piece for its cogent argument around user urgency. Stephen often reprises the argument too- here talking about boot times for Google Chrome OS devices. […]

  7. on personal metrics, talk to some triathletes to see a wide variety of toolkits.
    I have 3 different training logs. One is narrative to log emotional/life outside training/sickness, one counts the TSS (training stress score) of each workout and sums it over various periods, one is a vanilla log of workouts also summed in various ways.
    The trick is to reconcile the life-outside-training stresses with the training stresses.. unfortunately it’s not easily quantized..

  8. oh, and the vanilla log is also online, as a way of exposing the data to criticism 😉

  9. […] (BIRT, Jaspersoft, etc) – there’s a lot of room left for innovation. I’ve already written about the data that we’re all generating; the next step will be demystifying the process of […]

Leave a Reply

Your email address will not be published. Required fields are marked *