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.

Seeing as Max is Open Source, I see an opportunity here for some software to automate all this. Imagine dragging that downloaded zip file onto “MaxiTunes”, having it pop up a simple interface for entering metadata, and then it takes care of converting the tracks, adding the metadata, importing them into iTunes, and moving the original files to the trash. I’d be willing to contribute funds to a developer willing to give this a try.

Update: It turns out that I underestimated Max. Apparently if a FLAC file contains metadata, it’ll be preserved during the conversion. And if not, Max has built-in metadata editing. It can also automatically add converted tracks to iTunes, and even to a specific playlist. Impressive! I like Max a lot – it’s fast, effective, reliable, free, and Free. My only real quibble would be that the UI is a little confusing. Just a little though.

Your headline is misleading – I hope it’s an innocent mistake, as opposed to an intentional attempt to spread FUD about Open Source software. The crucial nuance that’s missing, both from the headline and the article, is that bugs such as this are just as likely to occur in closed-source software as in open-source — but with open-source software, they’re far more likely to be discovered, sooner, precisely because the code is available for anyone to peruse!

I expect more from Technology Review. Please consider adding a note or correction to this effect.

I’m not actually surprised or anything; TR has always had a corporate feel to it – it’s generally about MITers or ex-MITers making money, as opposed to pure excitement about technology.

Edit: Struck out the last paragraph; after Jason Pontin’s comment below, I realized that my impressions of TR are almost certainly out of date. BTW, he also responded to my comment; check out the discussion there.