I have been using a ChromeOs cr-48 device for a few weeks now, and these are my findings. I tried using the cr-48 for as many tasks as I could, including writing, image editing and even coding. Am I satisfied with the device? For some tasks, it’s just awesome. For other things, the entire cloud model is still not just there, either because of completely broken workflows, or small details.
For the tl;dr version, just jump to the last “Parting Words” paragraph.
Before we start, note that the opinions expressed here are mine alone and not necessarily match the ones of my employer. Also, if you quote any part of this article, please respect the page license (at the bottom of this page) and include an attribution notice leading back to the full article next to the quoted text.
Intro
I spend a good chunk of my life in front of a computer screen. I write blog posts, tweet, share photos, read news and huge Google Reader subscription lists, do the occasional creative work, like image editing and so on. Most of my time is spent coding, either for work or personal projects.
Also, I’m completely sold to the cloud computing model: I have something like 6 devices (between smartphones, tablets and PCs) that I use to access my data. Using local or 1-to-1 sync solutions simply does not scale (connecting a USB cable and syncing? You kidding me?). Most of my data lives in the cloud. I’d move to the cloud all of it if I could. Being a Google employee, I inherently trust their services and therefore most of my cloud lives off Google datacenters and services.
So ChromeOs and everything it stands for sounds like heaven to me. Does the real thing match my expectations? Some bits yes, some bits fail miserably.
Bear in mind that the standard I compare it against is my trusty 15’’ macbook pro and the standard computing day I usually have, so my usecase is about replacing a full featured laptop and everything I do on it daily with a cloud device. This might not be the main usecase that ChromeOs targets (more on this later), so keep this in mind while reading.
Also, when discussing the feasibility of living in the cloud, I mix issues that arise from the hardware I tested (which could be solved by switching device) with more fundamental problems in the cloud model. Please forgive the confusion, you’ll have to sort it out.
The good parts
Let’s start with the things that ChromeOs does really well.
Portability and life on the go: The cr-48 really shines. It’s small, compact, feels robust, lightweight, not so much bigger than a tablet (neither won’t fit in your pocket anyway) but definitely slimmer and smaller than my 15’’ macbook pro. It’s a joy to carry around. The devices that will actually go on sale on June 15th (announced at Google I/O 2011) look even better.
Battery life is very good. Haven’t measured it precisely, but I can use the cr-48 for hours at a time, and between usages, it sits on standby power for days at a time. In this respect, it doesn’t behave any worse than my macbook. On the contrary, it feels even better as a device to carry around.
Speaking of being on the go, 3G was not activated on my device, so I went for wifi only. If you happen to have an android smartphone, wifi tethering is just magic. Having an android phone with a 3G plan and wifi tethering enabled on it means that any other device you might have is connected to the internet wherever you are (or at least wherever you have 3G signal…). In my case, Nexus S + cr-48, means I have access to my cloud pretty much anywhere anytime. Your mileage may vary if you happen to live in areas of poor coverage.
The first time you try Google Cloud Print you will just want to cry for how beautiful it is. It just works: bring your cr-48 along, write some Google doc, print dialog, find a local printer via cloud print and you’re done.
Chrome sync (for syncing bookmarks, extensions, passwords and whatnot across all my chrome installations on different machines) is another of those things that just work and you don’t even realize how good it is until you miss it.
Text editing in the cloud. Just do yourself a favor and install SourceKit. Best cloud text editor out there, that syncs directly with DropBox. It is that good, and half of the ChromeOs experience comes from having the right extensions installed on the machine.
And finally, the hassle-free mindset of not being worried about losing the laptop, throwing it in a river, losing my data due to harddrive failure is just the extra cherry on top. It’s an expendable device (although I say so just because I got this for free).
Meh
Ok, let’s start with the things that are not that good, but I can live with. First, only one usb port. Ok, I could carry around a hub, but then I don’t want to have around a bunch of cables and/or dongles anytime I need to connect my mouse and my camera at the same time. Kinda goes against my previous statement of not wanting to connect USB cables to sync stuff, but until my camera improves, I’m stuck with it (I’m actually already using an awesome eye-fi SD card , highly recommended, but the eye-fi suite is not available for ChromeOs).
The keyboard is good. I really like the tactile feedback (my comparison is the Das Keyboard). However, it has no backlight. It’s one of those details that I really miss from the mac, and ultimately make macs stand out from the crowd (the other being the magsafe adapter). After you’re used to a backlit keyboard, the first time you try working at late night feels… strange. You actually need some other light source. Meh.
Performance is borderline: youtube is fine, youtube hd not so much. Videoconferencing over gmail is doable, but the fps is a bit low (my parents would not be able to track all the jumping and running my daughter performs in front of the camera). ymmv depending on which device you buy. Anything which is not flash-based performs reasonably enough though, including having half-a-ton of tabs open.
The LCD Screen feels a bit hazy or cloudy. Can’t say what it is (maybe the black is not so black, the color balance or whatever), but you definitely recognize it being less crisp than a macbook. Yeah, I know, two completely different price ranges and you have to reduce costs somewhere. I can still work in front of it for 6+ hours straight, so no big issue.
There is no remote to use with the cr-48. Minor detail if you want to connect the cr-48 to your TV and drive it from your sofa, while using the BBC iPlayer. Workarounds exist.
Profile switching is good: you can have as many profiles bound to your chromebook. I personally have several: my gmail account, my google.com work account, and a few other google apps accounts. Switching is easy, but switching speed should be improved. I’d like user switching to be as fast as alt-tabbing, but it still takes a good 5-6 seconds to get to the login screen after logout.
File management is ok-ish. I didn’t really play much with it, apart from connecting a usb key and verifying that you could browse the contents (what you do with them, apart from uploading them to the cloud, is another matter).
You are doing it wrong
Ok, first things first. The trackpad, oh the trackpad, is still the single biggest reason why I wouldn’t recommend the device. This has nothing to do with cloud computing, but it is so annoying that it almost makes the cr-48 unusable (although to be fair it greatly improved since the early days). Nothing new here, since half of the world already complained about the cr48 trackpad. Just plug in an external mouse (oops, you just lost your only usb port).
Most of the issues I have come from specific tasks and workflows that I tried to accomplish on the cr48, so let’s focus on each specific case. Mind that I consider the device good enough for content consumption, so let’s focus on content creation tasks.
Writing piles of text
No big deal, fire up Google Docs, your blog post editor or any text editor of your choice (SourceKit, SourceKit) and you’ll be good to go. The cr48 works just fine. This blog post was entirely written and composed on the cr48.
Basic image editing
Kinda works, I used Picnik for most of my needs. I don’t particularly like it (if anything just because it’s a just flash application) but it does the job. I tried a few basic things like the usual resizing, cropping, fixing color temperatures and such. Minor annoyance that only the basic downloader works with ChromeOs, which forces you to restart Picnik each time you download a photo, and it takes time. Picnik is quite slow to load (same goes for other cloud-based image editors I tried) and this compares unfavourably to the time it takes me to open Preview on the mac and resize an image.
I tried also something more complicated, like creating a party brochure. I ended up doing it in html5+css3, so the experience is the same you can read in the ‘Coding’ paragraph.
Some sharing tasks
The landscape is too varied here. Case in point: a friend of mine shared one of his Picasa albums with a few dozens pictures privately to me. I want to create a personal copy of said images. On my laptop I’d just hit Download to Picasa and I’d get the entire album in a short while; I’d then be able to reshare it as I like.
On ChromeOs I’d like to do the same. Lacking a decent filesystem interface, the closest equivalent would be to copy my friend’s Picasa album into a Picasa album of mine. Unfortunately, there is no way to do that (afaik), hence I’m stuck with accessing the photos through my friend’s album, hoping he’ll never delete them.
This is just a specific case, but it highlights the many inconsistencies that arise in the current crop of cloud apps when you try to use them for real, 100% in the cloud with no local fallback.
Presentations
So I tried using the cr48 to present my Google I/O html5 talk. The trackpad was so bad that it ruined the whole experience during the rehearsals when I tried to use the cr48, making some of them completely laughable. In the end I presented using my macbook. Trackpad aside, hosting a presentation should work just fine. Creating it on the cr48 is a different matter (see next point).
Coding
Ok, coding in the cloud is still a nightmare. I tried Cloud9 IDE first. Works fine until you are writing text. Then you want to attach an image, but cloud9 only supports drag’n’drop from your desktop (really, no ‘upload’ button ? or did I just miss it?). But there’s no desktop in ChromeOs, so you have to drag’n’drop from another tab where you’ve opened the file manager. But unfortunately that doesn’t work. Not being able to upload static resources to my projects is kind of a deal breaker to me.
So I switched to the other alternative: ssh-ing to another machine and using VI over there. Feels kinda cheating, since I’m using the cr48 only as an ssh client, but I guess it’s still fair game (and it recently improved thanks to the support of key-based authentication ).
So far so good until you want to attach an image or other static resource to the project you are working on remotely. Now you don’t have any local image editor, so either you already have the perfect image somewhere else, or you’ll have to: find the image on the web, have Picnik fetch it, retouch, refine, resize until you are happy, share the image on a publicly accessible location (such as a public Picasa album), Ctrl-C the public image url, ssh to the host you are using for development, wget the image using 3-finger click to paste the url to fetch.
All of the above to basically attach a static resource to a coding project you’re working on. No way, that’s too much of a productivity hit to be acceptable.
(alternative workflow: save image locally, upload to Picnik, edit, save edited image locally, upload to Google Docs, share with the world, get sharing URL, ssh to remote host, wget image from remote host via sharing URL).
A few improvements to the process can be made, but it’s still way too annoying compared to the simplicity of doing the equivalent task locally on any PC.
A coding task I tried was to write a simple Chrome extension, which sounds like something very popular in a world full of ChromeOs devices. You’d think it’s doable until you realize that you need a Chrome browser with access to the local filesystem to either package the extension or load it unpacked for testing. Oops. And if you develop the extension on a remote ssh host, you have to rely on community-provided scripts. Ok, still doable, but so much more painful compared to the alternative workflow on a standard PC.
Parting words
So what’s the verdict? Usability issues with the specific hardware of the cr48 aside (burn the trackpad), I think we are very close to having a completely cloud-based computing model, but still not quite there. The majority of hurdles are small annoying details that you figure out only when you try to use the cloud fully, with no local fallback, and therefore will hopefully be solved in a few iterations as soon as cloud-based products realize they are not really entirely cloud-bound (like the Picasa album copying case I described). There are still a few areas where a full cloud model is still a long way to go: any serious coding, especially if you target non-web languages for which you need a cloud-based compiler. Likely any serious creative work (painting, 3d modeling, professional photo editing) is still out of cloud reach.
Content consumption has already transitioned to the cloud big time. Content creation is already there for most small tasks like maintaining this blog. Still takes some work to be self-consistent (I want to use cloud-based tools to produce and ship cloud-based tools), but we’re getting there.
As a personal experience, I find carrying around the cr48 more pleasing (in terms of things I can do and how productive I feel) than carrying around my motorola xoom (where content creation feels mostly impossible, aside from the occasional email writing), so I definitely see a market opportunity for ChromeOs. My personal ratio of cloud adoption is probably higher than the average consumer, so my perception of its usefulness is somewhat skewed.
Still, rather than selling this as a device targeted at the general consumer, I see the enterprise/academia adoption as the killer usecase: as an IT head admin for an enterprise or campus, I’d be extremely happy to hand over expendable, cloud-based, cloud-configured devices to all the staff, especially if Google provisions and replaces the devices as announced at Google I/O 2011. As a staff member, I’d be equally happy to have all my content readily available on any of my colleagues Chromebook whenever I forget my own chromebook at home and/or I lose it for whatever reason.
Go ChromeOs.
