Student, geek, unicyclist and more

Today, Google held a press conference regarding Google Chrome OS, as well as releasing a slew of videos pertaining to it. They’ve discussed the major benefits, and they come quite clear: less cruft, less software maintenance, less waiting and quick access to powerful web-apps for day-to-day use. Along with Google Gears support for offline use of these web-apps, this makes for a unique, and somewhat appealing OS. However, what major flaws could be presented in this new approach to OS design?

Problem Numero Uno: Consistency

At the moment, regardless of which operating system you own, there’s some consistency in how it looks and feels: windows look and feel the same, things like menus are delivered in a consistent manner, opening a file is done the same way, settings are changed in a consistent manner, right-click behaviour is the same, a button looks like a button looks like a button, et cetera so on and so forth. However, this isn’t true on the web: instant messaging is different between, say, Google Talk and Facebook, Hotmail and GMail are vastly different, and Photoshop.com doesn’t share anything at all with Google Docs. All these applications behave differently, have different ways to change settings, different menus, different layouts, different colour schemes (and different ways to change the themes, or none at all)… all this stuff would behave the same on a full desktop OS (Menus, for example), and a lot of it would follow the rest of the OS (scheme and theme).

The problem is: the web was not designed to be an application framework, but some people are treating it like it is and some people are not treating it like it is (I treat it as a source for information, myself).

Problem 2: Place

Another problem with the browser-based OS is: where does everything go? The obvious answer is, “the web browser”. In this case, the obvious answer is right, and therein lies the problem.
I have on my laptop an application dedicated to microblogging. It’s better then the browser-based Twitter and identi.ca clients in that they’re both in the same window, I can update both and read both easily, as well as quickly seeing updates without needing the window open. Moreover, it fits in nicely with the rest of my desktop and even includes features like announcing what song I’m listening to. The best bit, however, is that I don’t have to give it my full attention for it to be useful.
Likewise, I have an instant messenger client that connects to Windows Live Messenger, Facebook chat, Google chat, Yahoo and Skype. It gives me the ability to group contacts into, “people”, so if I have the same person e.g. on Facebook and on Windows Live, I click and drag one on top of the other and it makes them one, “person”, preferring Windows Live over Facebook by default but letting me message them and receive messages from them on Facebook if that’s all they’re online on. It also looks and feels like the rest of my desktop, and also includes such features as Now Listening, contact online/offline statistics and, finally, lets me see messages without having the window open. Again, I don’t have to give it my full attention for it to be useful.
“Ah,” you might be thinking, “But Web applications could be made that do that, too!”
Well, that comes at a price: to use them, I need to open them in a new, full-screen tab or window. To see if anything interesting’s gone on in them, I have to switch to that tab. Apart from being completely separate, and as show above potentially very inconsistent, there’s currently little or no way to find out e.g. what Youtube video you’re watching, what Magnatune song you’re listening to (since an iTunes client won’t be available for Chrome OS). There’s no feasible way to implement, “Now listening”. There’s no way to make them informative, but not distracting. There’s no way to make them look and feel like the rest of my web-apps, like Google Docs, Hotmail, Youtube etc. They’re separate, unique, potentially confusing, more difficult to create in the first place (if you’re not going to cheat and use Flash) and finally, just not integrated.

Maybe all this will change in the future, but let’s not forget: the web was designed for information, not application development. (KDE propaganda ahead) My personal opinion is that the web browser shouldn’t become the desktop, but that the desktop should integrate tightly with the internet, and that desktop applications should integrate with each other to better make sense of the information they grab and provide extended features and functionality. This is quite on-par with the KDE 4 vision, and that’s why KDE 4 appeals to me so much. (end KDE propaganda)

Problem 3: Reliability

Lots of people have come talking to me about the potential problem of reliability, mainly regarding internet connection (“What happens when you disconnect while writing a document?”). However, that problem is not the one I wish to talk about right now – instead, I’m going to talk about the problem of changes to applications.

In this sense, Google’s touted strength (“You don’t need to upgrade anything any more!”) is also a fundamental weakness: you no longer choose whether you use the old or the new version of a given application. Why would you want to, anyway? You’re always going to want the latest version, right? Anyone still working on Microsoft Office 2003 or Windows XP will likely understand.

When applications get updated, things can change. Sometimes things change very dramatically (Office 2003 –> 2007, KDE 3.5 –> KDE 4.0, Windows XP –> Windows Vista etc.), sometimes things break very badly. It might be something like not being able to open old documents, new bugs and crashes, re-designs in the interface, new hardware requirements, new SOFTWARE requirements et cetera so on and so forth. On your average computer, though, you can decide not to upgrade: you could, if you wanted to, use Window 95 until way after 2010. That’s a bad example, though, so I’ll choose something more real: you could use Windows XP after 2010, or Microsoft Office 2003, or the old Windows Messenger. You could decide to use KDE 3.5 until after 2020, if you really wanted to. You could even install Office 2003 and 2007 side-by-side, and alternate between them, or you could use two versions of Windows side-by-side as well. Importantly, the point I’m getting at is that you can decide that an upgrade isn’t going to be as good as the thing you’ve already got.

Chrome OS, on the other hand, takes that decision away from you: imagine if Google(/Microsoft) drastically changed their on-line office suite. If you don’t like the new one… well, tough. If it breaks compatibility with your old documents, or hell, if it breaks compatibility with documents other people send you, then sucks to be you, but hey! At least you didn’t have to upgrade, right?

The fact is, not only do users make these choices all the time, but so do computer manufacturers. Computer manufacturers also have a tendency to make, for example, several versions of Windows available on the same computer (a choice of one, not all of them generally), and this preserves your ability to choose which you prefer. With Chrome OS, that all magically disappears, and we all have to hope that either the developers never break anything (which won’t happen), or otherwise make several versions of their products available at the same time.