My Life on Chrome OS
My main laptop is a Pixelbook. This is the second time in the last 5 years that I’ve used Chrome OS over multiple months on a daily basis. In this post I iterate over a few thoughts on the state of Chrome OS and where I hope to see it in the future.
The concept of Chrome OS is a little strange at first. An operating system that is a browser. Of course, there is a lot more to an OS than just a browser. Bootloader, hardware drivers, windows management, network connectivity, security, memory management etc. In fact, another way to think of Chrome OS is Linux with a Chrome-based UI.
In this context, one could argue that Chrome OS might be the most successful Linux distribution out there. It’s a spurious argument perhaps, but it gives us a hint at some of the potential for Chrome OS.
Whilst Chrome OS isn’t mainstream, it also isn’t new anymore. At time of writing I’m on version 69.0.3497.95. This means we shouldn’t give Chrome OS too many free passes at this stage.
Apps or Browser or Both
After several years away from Chrome OS, it feels like there is still a problem with the UI concept.
Essentially there are five different ways in which windows exist on Chrome OS.
- Tabs in a single browser window.
- Multiple browser windows (with/without multiple tabs)
- Separate standalone windows that resemble apps rather than browser tabs.
- Android apps which run in their own window along with Android-style up navigation.
- And now Linux apps (for Chrome OS 69 on Pixelbooks)
This wouldn’t be so bad if you could pick one approach and stick with that. Unfortunately Chrome OS constantly drives you back to browser tabs.
For example, I have YouTube set to open as its own window so it feels like an app when I open it from the launcher. However, if I click on a YouTube URL, it gets opened in a browser tab. Now I have YouTube scattered across app windows and browser tabs. What’s worse, I could also have YouTube running as an Android app.
Here are some options for how this situation could be improved.
Enforce a browser model
Chrome OS could drop the concept of opening browser content as it’s own window and just stick with browser tabs.
This has the benefit of being consistent with how Chrome works as a browser on OSX or Windows. I also know what to expect when I click on any URL, even if there is an ‘app’ in the a Chrome OS launcher.
The downside is that it no longer becomes possible to group windows belonging to the same app — something that most operating system UIs would do. If I am editing multiple Google Docs, I would have to manually make sure they are in adjacent Chrome browser tabs if I wanted them grouped together so I can, for example, easily jump between related docs whilst editing.
Enforce a window model
Alternatively Chrome OS could say everything opens as it’s own window without the browser chrome. So, five open Google Docs means five open windows.
The advantage is that apps will feel like apps. Open a link, and a new window would open. Click on a document in Google Drive and a new window will open.
The downside is that this feels like a step backwards to the world before browser tabs. Personally, if I didn’t have tabs, I would have so many windows open it would be painful. So, this breaks the existing model of browser tabs but fixes the problem that apps should open as their own window.
Apps vs web pages
A simple model, consistent with other OS UIs, is to enforce that anything with a launcher icon always opens in a new window — irrespective if the launcher icon was clicked or a URL to content for that app was clicked.
Items with launcher icons become apps — just like MS Word is an app on Windows. The Chrome browser itself would just be one app.
Borrowing from the concept of Android Intents, apps could register that they are possible consumers of specific types of URL. Send me a link to a Google Doc on Slack? It opens in the Google Docs app in a new window.
The added benefit is that Android apps could also register themselves and launch based on URL clicks — something that doesn’t happen today on Chrome OS.
Inconsistency, inconsistency, IncOnsIstEncy
Chrome OS is littered with inconsistencies which really contributes to the sense that the OS isn’t a serious player. These aren’t a serious as the browser vs app inconsistencies mentioned above, but they do add up.
Here are the ones that bug me.
- Inconsistent keyboard shortcuts. For example, CTRL + Enter in the file manager will allow for file name to be edited. The same shortcut on Google Drive (also a file manager) does nothing.
- Switching keyboard languages (CTRL + Space) work in the browser window but not in a standalone window.
- Open a PDF in the File Manager and you get a separate browser tab. In Drive it’s an in-line viewer on top of the Drive window.
- Effectively there are 2 file managers. File Manager mounts Google Drive as a device. It’s a nice touch but Google Drive is available in the Chrome OS launcher as well — it’s more feature rich too.
- Pinning vs Shortcuts. I can pin a tab in a Chrome browser but also create a shortcut in my Chrome OS launcher. These are similar concepts but there isn’t a clear rationale for using one vs the other.
- Too many app concepts. I have a mix of Chrome Extensions (e.g. Pocket or Pushbullet) , Chrome OS launcher icons (e.g. Drive, Google Docs) and Android Apps (e.g. MS Word). Since Chrome OS 69 I now have Linux apps.
- Android apps and their web siblings sometimes work differently. Maybe this isn’t a Chrome OS problem, but supporting two app runtimes means users are exposed to the differences. An example would be Dropbox Paper which I use to write my articles. I use the Android app for offline editing. But it is missing a bunch of features found in the browser version, such as viewing editing history, ‘templatising’ a doc or exporting to another format. I’m not knocking Paper…I love it and use it every day (although offline editing in the browser would be awesome). But the inconsistency between Android and web is part of my Chrome OS experience.
- Android apps as default. Say I want Word documents to be opened in the Word Android app by default. That doesn’t seem to be possible and I have to right click on a downloaded docx file and select ‘Open with…’ every time.
- Printers are in advanced settings. This isn’t a big nuisance but seems silly to hide printer settings towards the bottom of the advanced settings rather than as part of the device settings, like monitors. Part of me applauds Google by de-prioritising printing…but at the same time connecting a printer shouldn’t be and ‘advanced’ setting.
These inconsistencies need to be addressed, and contribute to Chrome OS’s identity crisis. Looking at the state of the UI on most Linux distros will quickly tell you that inconsistency kills the user experience.
Compatibility
Most of the people I work with use OSX, some clients use Windows. Nobody else I know uses Chrome OS.
Luckily a lot of our work is done via the Google Docs suite and web-based apps like Jira, Slack, Zeplin and others. I rarely come across situations where I’m stuck. Google have long since figured out how to support MS Office file formats.
Apple Keynote is the notable exception. I still get a fair number of these sent to me. Generally Keynote feels alien to me anyway, so I’m not sad to have to request a pptx version instead.
However, I do miss some OSX apps. For example, Mindenode — I ended up going with Mindmeister as an online alternative. Simple image editing feels like it is missing. I used Acorn on OSX. I haven’t found a nice alternative. Sumopaint and the Snapseed Android app acts as a substitute.
PDF support compared to OSX feels very week. This is one area I would like to see improvement, given Chrome OS should otherwise be great in business settings.
More unusual file formats like 7Z compresses archives are a pain. It’s rare to get these and with the Linux VM on Chrome OS I should be able to open this type of file now anyway.
I’m pleasantly surprised how well video plays on Chrome OS. It’s not a big part of my usage, but I haven’t had a problem with any video yet.
Offline Support
For document editing, Google has offline mode working nicely on Google Docs, Sheets, Slides and Drawings.
However this is another point of inconsistency. I have access to offline version of Google Docs, Sheets, Slides and Drawings, but not for other files like PDFs that are in Google Drive.
The issue is that Chrome OS is based on the concept of cloud-based document storage. In contrast to Dropbox type models which are based on syncing local files, on Chrome OS files are only stored in the cloud.
(You get a Downloads folder on the device, but you don’t otherwise have access to the local file system.)
Surprisingly, offline support for Gmail isn’t good. You need to use the Gmail offline chrome ‘app’. The good news is that this is finally being solved and offline support will become part of Gmail.
Sometimes I will use an Android app for offline work. Dropbox Paper which I mentioned earlier is one example.
I have been travelling weekly over the last months. I haven’t found a huge issue with Chrome OS’s offline support. A lot of my blog writing is done in a plane at 10,000 meters above the ground.
Peripherals
I’ve been impressed with how well Chrome OS deals with peripherals.
I’ve connected to a range of HDPI monitors and they seem to generally work well. External displays are consistently recognised and config options offered. Colours do vary across monitors, but I’ve found that issue with all OSs. Projectors have also been working seamlessly, both with USB-C to HDMI and VGA.
I haven’t had a problem with printers either — although I’ve only got 2 configured. Back in 2013 only Google Cloud Print was supported. Today you can connect direct to printers over Internet Printing Protocol and others. What I really miss though is the ability to use my office printer’s scanning from Chrome OS. For business users this sort of thing should be possible. (OSX is great on this front.)
Mouse connections must be done via Bluetooth. Annoyingly my Logitech mouse uses it’s proprietary wireless protocol and I don’t have a USB-C dongle. Might be something I could get, but I dislike that setup anyway. My Apple mouse seems to work well.
I have no experience with cameras or other peripherals. My guess is that this would be quite hit/miss.
I haven’t used the Pixelbook Pen — I keep meaning to get one. Given I don’t use the tablet mode too much, maybe I’m not missing much.
Performance
I am lucky enough to have the top-end Pixelbook. It’s an i7 processor with 16GB of RAM and a 512GB SSD. I don’t experience any performance problems.
Frankly, for the price of the top spec this should be a given. Over time I am not seeing any performance degradation either which is pleasing.
Recently with the ability to run a Linux VM and install Debian packages, I was curious about performance. I didn’t need to worry. Even heavy apps like Android Studio run well. (It’s early days with Linux apps — maybe it’s premature to make final judgements.)
Battery life on my device seems OK but nothing spectacular.
There are many lower-end devices out there. I don’t know how well or badly Chrome OS performs on these but I can imagine it being quite varied.
Stability
Back around 2013 my Chrome OS would regularly reboot for no reason. That is a thing of the past. I haven’t had any OS crashes so far.
Android apps run fine and I haven’t see these crash much either.
Bluetooth however does get disabled regularly. Restarting the OS is the only way to enable it. Given Smart Lock is based on Bluetooth, this is a real nuisance. The high price of my device should mean this isn’t a problem.
I haven’t seen this since installing Chrome OS 69 — maybe it’s fixed.
Tablet mode
The Pixelbook can run in tablet mode. Flip the screen 180 degrees and the physical keyboard is replaced by a soft keyboard. It works but the hardware solution feels half-baked.
Chrome OS doesn’t feel particularly good or bad as a tablet OS. This is expected to change with the next release, version 70, and Chrome OS will have tablet-friendly features. How good that experience will be is something we will have to see.
Personally I am not a big tablet user — at home I have a Kindle fire for reading which works well for me. The Chromebook feels too large & heavy to be a comfortable tablet.
Chrome OS for developers
With Chrome OS 69, Pixelbook owners can now enable a Linux VM and install Debian packages. This feature is still marked as Beta — it hasn’t made it to other Chromebooks yet.
I’ve been running this for a couple of weeks and I’m feeling confident that my Pixelbook could be a good development laptop. I’m currently edging back into developing after a few years off, so I think a different assessment might be reached by others.
However, Android Studio runs fine and Android/Flutter apps can be run directly without an emulator. (Remember Android apps can be installed on Chrome OS…)
With the Linux VM, you can now use the terminal as you would wish. Integration is done well, with a mount-point for the Linux VM in the file manager, launcher icons for Linux apps and Chrome OS styling consistently applied to title bars.
What’s in store for Chrome OS?
The next version of Chrome OS, version 70, is slated to be a big update. There are lots of rumours that Google will be releasing a detachable tablet/keyboard device and several of these updates coincide with a new hardware release.
I think there needs to be some soul searching about what Chrome OS is and what Google want it to be. Student-friendly cheap laptop? Business machine for companies using G-Suite? Developer machine?
The browser-as-an-OS concept was intriguing but overall the experience remains messy and inconsistent. I hope that a clearer approach is on the way.
It remains to be seen whether the Fuchsia project will ultimately be Google’s future OS. I wouldn’t put it past Google to replace Chrome OS at some point. This is probably a few years away so there is plenty of time to improve Chrome OS.
TL;DR
I first used Chrome OS in 2013. It was early days and I was experimenting with a Chromebook as my main laptop. In the end I went back to OSX.
Fast forward to 2018 and I’m back on Chrome OS with a Pixelbook as my main machine. Over the last 5 years Chrome OS has remained pretty similar. I noticed some UI improvements and better support for peripherals but the experience hasn’t changed drastically.
Chrome OS still feels like it has and identify crisis. Being able to run Android apps, and now Linux apps, will only make this situation worse. At some point, Google needs to address the uncomfortable role that the browser plays in Chrome OS. The Fuchsia project could change all this by offering a Chrome OS replacement.
By far the most exciting developments on Chrome OS is the support for a Linux VM. This could open the door to Chrome devices being used for software development — I know a few people itching to swap their Macbook Pro’s for a Pixelbook.
