I’m looking forward to Chrome OS. It promises to be radically faster and simpler than all of the currently popular operating systems. I’m hoping it actually delivers on that promise, I think it’ll be good for computing!

OK, I was going to end the post there, but my (business) partner Tim just informed that I do have some insight and analysis I can throw into the ring. He asked me: if Google’s vision of the future of computing is the web, and they already have their own browser, why would they build their own OS? Aren’t browsers the gateway to the web? Here’s how I replied:

The key to understanding this is in the name of the browser and the OS: “Chrome.” In computing, “chrome” is “the borders and widgets that frame the content part of a window” (Wikipedia) — in other words, it’s the stuff around the stuff that really matters — the “content” — the document, messages, information, movie, pictures, etc that you’re reading, writing, watching, editing, or otherwise interacting with. It’s become a common wisdom in user interface design that chrome must be minimized, to the point where it’s invisible, if possible. (I suspect Edward Tufte has been one of the supporters of that philosophy, but I don’t have any sources handy.)

With that understanding, Google’s philosophy seems clear to me. By naming their browser, and now their OS, after the stuff that doesn’t really matter, they’re revealing their thinking about what does really matter: the web. So this new OS is about delivering the experience of the web unshackled by the complexity that the modern operating system is weighed down by.

I apologize, but to continue I need to take a moment to talk about what the term “operating system” means. There have always been different definitions of the term. To many, it’s all the stuff that’s included with your average copy of Windows, OS X, or a Linux distribution: a file manager, notepad, calculator, control panel, web browser, email program, etc, etc, etc. But technically, those are all actually just applications which are bundled along with the OS; the OS itself is the code which manages the hardware, and provides an abstract interface to those applications, so they themselves don’t need to worry about the vagaries of hardware. But since the term OS has fractured over time, a new term arose to describe that core part of an OS: it’s called the kernel. A modern kernel, along with integrated extensions and drivers, handles important things like hardware resource allocation, memory management, power management, networking, and all the other low-level tasks which are kinda boring but also critical to an effective computing experience.

Sundar Pichai, Google’s VP of Product Management and Linus Upson, its Engineering Director, wrote in their announcement: “The software architecture is simple — Google Chrome running within a new windowing system on top of a Linux kernel.” This sounds innocuous enough, but it’s really a radical challenge to all the mainstream operating systems available today. Windows, OS X, and the mainstream Linux distributions all have a kind of a kitchen sink approach; they include a kernel, and a windowing system, but they all also include much, much, much more — complex file systems, file managers, outmoded and ineffective security models, thousands of device drivers for devices most people never use, and hundreds and hundreds of bundled applications — they’re trying to be all things for all people. Chrome OS is a challenge to that entire model, and if Google manages to pull it off, it may be an end to there being a single dominant model for how end-user “desktop” computers work. (By “desktop” I mean any and all descendants of the typewriter-based terminal model, centered around a keyboard, including laptops.)

Essentially, the philosophy behind Chrome OS is for it to contain only the bare minimum of systems necessary to boot up the hardware, connect it to the network, and load up the web. Just about everything else just won’t be there. Sure, that won’t work for some people — but my guess is that it will work for a whole bunch more. And by radically simplifying the software stack, that stack can be radically faster, more reliable, and more secure. Sounds worth a try to me.

When I change an event, I should be able to enter a note to specifically explain the change itself, which would be included in the update notification email. Right now, we put notes in the Description field, but it’s confusing.

At my org, we always try to allow Guests to modify the event. That way we can more collaboratively adjust our shared schedules. But we frequently forget to check it. It’d be very helpful for us if there was a setting to make that option checked by default for new events.

GooMaps told me to take the B train today, but it’s not running today,
due to Thanksgiving! If I’m going to trust Google’s transit
directions, it needs to account for all the complexities of the
transit system schedule, including holidays!

It got me thinking, so I posted this comment:

Good points, Rich. I’d like to note that there’s a related issue: backup. Most people don’t backup their data, even that which is crucial to them. I think the basic reason for that is that it’s always been way to complicated to do so. Apple’s Time Machine offers some hope in that area; it boils backup down to its essence, and hides the complexity, so average people can backup without too many hassles. And that’s great. But consider that it took twenty-four years — from the release of the first Mac to that of OS X 10.5 Leopard with Time Machine — for us to come up with a simple, reliable, effective, and affordable solution to desktop backup for average people — and it’s still not built-in to the most common OS. And — the irony — just when we’re finally on the cusp of slaying that dragon — at least for Mac users — the game goes ahead and changes! Now people are keeping their data off of their desktops, on other “people’s” computers! Now we have this entirely new dimension to the problem, and once again, it’s being approached haphazardly, obliquely, and without coordination.

You focus here (as you frequently do) on the divide between the web and the desktop. That’s one way of looking at the problem. But if you look at it from the perspective of backup, you could re-frame the problem as a rule: important data should always exist in more than one place. And by “place”, I mean “responsible party” — I’m sure Google has a strong backup plan, but if they hold the only copy of one’s data, and their system goes down, that means that the data is at least unavailable for some period of time.

I try to follow this rule, which is why most of my “cloud” data is mirrored on my laptop, and I’ve got multiple geo-redundant backups of my laptop. I set up a desktop mail client to synchronize with my Gmail accounts using IMAP, which means I have not only a backup of my mail, but a local working copy — if Gmail goes down, I can still read and send mail.

The thing is, that’s all well and good for me, but I’m a geek. I have no illusions that we can expect average people to set up or use such a setup. So I agree with you that it’s time for the cloud service providers to step up and try to address this problem, on the behalf of their users. Gmail, for example, should provide a “Gmail Desktop” application, which would store all of a user’s mail locally and synchronize with the service. When on their own computer, a user could use the app, which would automatically provide the benefits of backup and offline use, on the fly, in the background; when away from their own computer they could still, of course, use the web interface.

Thanks for the interesting thoughts!