Test-driving Chrome for Ubuntu
With an alpha version of Google’s Chrome web browser recently released, I’ve been using it on Ubuntu for a few days. Below are some thoughts on the new browser and its ability to improve the Ubuntu experience.
Before delving into an evaluation of Chrome for Linux, however, I should point out that the version I tested (which I installed using the 32-bit Debian package available here) was very explicitly advertised as unstable–so much so that a popup window appeared warning me to expect strange behavior the first time I launched the browser. For this reason, I’ve assumed that the annoying bugs I’ve encountered in Chrome over the last few days will be fixed in due time, and haven’t mentioned them below.
I should also make clear that my chief basis of comparison for my Chrome experience is Firefox 3, which I’ve used for the last several years, and which is the default browser in modern releases of Ubuntu.
With that out of the way, here’s a run-down of what I liked and disliked in Chrome running on Ubuntu:
- simplified interface – Firefox’s interface is not bad, but Chrome raises the bar by removing extraneous menus by default in order to maximize the display area of web pages. I first found this frustrating, but quickly realized that I can get along very well without Firefox’s extensive menu. Chrome’s web-history and download-management interfaces, which are rendered as an ordinary web page in their own tabs, are also more convenient than the Firefox equivalents.
- faster ‘awesome bar’ – the awesome bar in Firefox 3 is a great resource for locating web pages quickly, but its tendency to lock up for a few seconds while trying to process a query can be frustrating. I have yet to experience equivalent behavior with Chrome’s implementation of the awesome bar. Perhaps I simply haven’t acquired a large enough web history to bog the utility down, but given Chrome’s focus on efficiency, I’m hoping that its awesome bar will indeed prove as zippy as the rest of the application.
- resource management – not surprisingly given the observations above, Chrome uses substantially fewer resources overall than Firefox. With the same pages open in both browsers for the same amount of time, top reported the following for Chrome and Firefox:
Chrome’s CPU usage exceeded Firefox’s occasionally, but in general, Chrome proved a much more efficient browser. I especially appreciate its efficiency on my tired Pentium IV laptop, which has precious few megabytes of RAM and CPU cycles to spare.
- UI integration – although it’s written using the Gnome GIMP toolkit, which should make it fit seamlessly into Gnome, Chrome looks out of place on my desktop. I wouldn’t mention this if the Chrome developers hadn’t initially complained loudly about the lack of UI consistency on Linux, and reluctantly settled on GTK+ as the basis for their port. While I’m sure the final product can be made to integrate cleanly into the Ubuntu desktop without much trouble, I was surprised that the alpha release looked more like something out of KDE than Gnome.
- default search engine – unsurprisingly, Chrome’s built-in search engine defaults to Google. In the Windows version, this can be changed either during installation or by right-clicking in the address bar. Not so in Linux–if you prefer a different search engine, you’re apparently out of luck. Perhaps this will be addressed as Chrome’s Linux port matures, but if not, Google can expect a lot of flak from Ubuntu users complaining about their freedom of choice being denied.
Overall, I’m impressed with Chrome so far. Its tiny resource footprint is likely to score big points with Linux geeks who like their machines to run as efficiently as possible, and with users seeking a more responsive browser than the mainstream offerings. The current lack of integration into Gnome and the inability to change search engines (not to mention most other preferences) is discouraging, but we should spare final judgment on these issues until Chrome’s Linux port becomes stable.