Midori

A small and simple WebKit browser. I have been looking for a good WebKit browser to use in linux… unfortunately this isn’t it. The performance is pretty decent when loading and using the page itself, but the actual program UI is clunky, slow, and buggy. This one needs some serious GTK+ work before it can expect to be a real choice for Ubuntu users. It hasn’t crashed on me or anything, but the fact that the toolbar can get resized if you switch tabs or use the back/forward buttons should not happen, but it does and it happens slowly.

Epiphany

Uses the Gecko layout engine, like Firefox, and seems to do it a little slower than Firefox. Some of the tab behaviors are odd and difficult to configure. When I click on a link that should open in a new window I generally don’t want it to appear in a new window, I want it in a new tab. Back and forward history seem to be shared among tabs in the same window? I guess I could see how that would be useful, but it is very odd behavior. Double clicking in the empty area of the tab bar doesn’t spawn a new tab? Also odd. All-in-all this isn’t a bad browser, it just doesn’t seem to do anything better than Firefox, which I was really hoping for.

Galeon

This is another browser using the Gecko layout engine. This browser actually feels pretty fast. I have enjoyed using it. However, it is not without issues. On a netbook it is very frustrating that I cannot seem to configure the browser’s UI to take up a minimal amount of screen space. The stock back and forward buttons are huge, and the height of the toolbar they are on is way to large. Even in full screen, the UI takes up too much vertical screen real estate.

These are the only alternative browsers I have tried so far. None of these browsers offer a good alternative to Firefox on a screen-space-constrained netbook. Not a single browser here had a windowed or full screen mode that offered as much screen real-estate dedicated to displaying the web page as Firefox (when properly configured). Even after exploring the various configuration options there appears to be no way to make ANY of these browsers use less or even similar amounts of screen-space for their UI compared to Firefox (not counting the options to remove the toolbar entirely, which is not realistic) . On a screen this small page real-estate becomes very important. There should at least be OPTIONS to make the UI as minimal as possible, but the options provided are simply inadequate. This compounds the fact that Midori and Epiphany both feel slower than Firefox. Galeon, which generally feels as fast or faster than Firefox, is crippled for me by the fact that the minimum usable UI configuration is the largest of the group. Also, not a single one of these browsers offers the ability have a FULL full-screen browser, like Firefox, where the UI auto-hides and the web page is given nearly 100% of the screen. I hope with the growing number of netbook users and the growing number of them running linux this will have to be a form factor each brower’s community will develop for and offer more customizable/minimizable user interfaces.

I haven’t tested this in Greasemetal (Google Chrome) or Greasekit (Safari) yet, but it uses very simple and standard javascript DOM to work, so I’m pretty sure it will work fine in other userscript environments.

Pretty simple, but its something I really wanted so I figured I would share in case anyone else needs a tool like this. Also available through userscripts.org.

Let me know if you can think of any other sites this would be useful for.

Here is a quick overview of my thoughts on Google Chrome:

Pros

• Super fast rendering
• Super fast javascript and great inspection tools
• Super fast user interface
• Very compatible and standards compliant (thanks webkit!)
• Simple, clean, attractive, and intuitive UI
• Fantastic stability and resource management

Read on for more detailed thoughts about the Google Chrome beta browser.

Throughout this article I will mainly be comparing Google Chrome to Firefox 3.0 for a number of reasons. First, Internet Explorer is garbage and isn’t even trying to be competitive. Safari and Opera are both great browsers and worthy of a comparion, but to be honest I am not intimately familiar with either of those browsers and taking the time to familiarize myself would cause the scope of this article to balloon from a first impression to a full-on review. I have chosen Firefox 3.0 as a point of comparison both because of my familiarity with that browser (its the browser I use 98% of the time) and the fact that it is far and away the most popular open source browser at this time.

When you first start Google Chrome you are greeted with a pleasant start page wrapped in an even more pleasant interface. It has conveniently imported my bookmarks, history, settings, and other things from Firefox for me. How thoughtful.

The first thing you see when you start Google Chrome

User Interface, or a Welcome Lack There Of?

Google Chrome really hammers home Google’s minimalist approach to user interfaces. The number of pixels devoted to interface in Google Chrome is significantly less than in Firefox 3.0 (which itself is much less than Internet Explorer). Even if you hide the Bookmark Toolbar in Firefox, Google Chrome still spends significantly less screen real estate on browser interface than Firefox 3.0. Chrome also shaves some additional screen real estate by ditching the status bar. Instead, Chrome uses a status “overlay” that is overlayed across the bottom of the page only when there is useful information to show you.

You will find there are no File, Edit, View, or any other standard menus anywhere to be found in this application. They aren’t hidden like they are in Internet Explorer… they simply don’t exist. All operations are performed through one of the two menus in the top right of the interface. One looks like a page icon and is used to instatiate a new tab, new window, or to perform various operations related to the current page. The other menu is shaped like a wrench, and as you can probably guess is used to perform application level operations, like accessing Chrome’s option screens or opening browser history.

At first the address bar in Google Chrome seems much like the Awesome Bar in Firefox 3.0… it performs full text searches through all of your history and bookmarks as you type. However, if you look closely you’ll notice that Chrome’s address bar also throws in suggestions provided by Google Suggest and gives you easy access to perform web searches against your default search provider (doesn’t have to be Google).

This an other features are explained right away by Google Chrome through a welcome tab that is opened when you first start the browser. These brief feature descriptions are great, and you can easily click through for more detailed information including video walkthroughs and even a comic book.

Google Chrome Welcome Page

The next new feature Google Chrome brings to the table is its “incognito” mode. You can open a new window “incognito” and any browsing done in this window will not show up in Chrome’s history, will not be cached in Chrome’s cached, and any cookies will not be stored with the rest of your cookies. This won’t help you if someone is monitoring your network, but this should definitely help keep your browsing history away from your kids or a nosey roommate. Some might even call this a “porn” mode. I am sure there are many practical uses for this feature besides looking at porn… I can’t think of any right now, but I’m sure there are reasons.

Regardless, the incognito mode works great and does so in style. I really love the visual cues used on incognito windows and the great little page that explains what incognito does (and doesn’t do) when you first open one of these windows.

An "Incognito" Window Opened for the First Time

The final feature I will discuss is what I feel is the most important and most unique feature Google Chrome offers: robust process isolation and resource management. Every single tab opened in Chrome spawns an independent operating system process, which is independent from the main Chrome process and independent from processes for each plugin being used. Every time you close a tab the corresponding process is killed and every resource it was using, including every last byte of memory, is entirely reclaimed. Even the simple act of typing a new address in a current tab will result in the process and resources for that tab being trashed, and a new process will be spawned. This is an incredibly smart way to manage browser resources and after seeing it in its simple glory it seems so obvious now. I am truly shocked this has not been done before.

The benefits of this approach may not be immediately obvious, but consider this example. You are using the Adobe Acrobat Reader plugin in one tab and it crashes. Normally this will crash the entire browser, but not with Chrome. Google Chome has isolated the resources of that tab so well that even if the “browser” crashes the result isn’t what you might think. Google Chrome isn’t just a web browser, its more of a browser manager. Each tab is isolated in such a way that it is its own browser, so if it crashes the remaining tabs will be completely unaffected and Google Chrome will keep running like nothing has happened. To make this scenario even easier to deal with Google Chrome includes its own task manager which you can use to manage each process (which can be a plugin, a window, or a tab). Should one of these processes stop behaving properly you can kill it, just like you would kill a crashed or stubborn program the task or process manager of any operating system.

Chrome's Task Manager

As you can see from the screenshot above, Chrome’s task manager also includes a link titled as “Stats for nerds”, which opens a tab that provides detailed resource usage information and even compares the resource usage to any other web browsers running on your system. This information isn’t incredibly useful, especially not to the average person… but Google is right about one thing: nerds love stats. I appreciate the nod from Google… even if the realization that I find statistics on memory usage entertaining hurts my self esteem just a little.

Performace: Subjective Looks at Speed and Memory

I haven’t set up any scientific benchmarks. I haven’t tested for memory leaks. I haven’t even done any hard number crunching over a vast number of websites. I have just used Google Chrome and occasionally checked its memory usage and compared it to Firefox 3.0. First, lets talk about speed…

Firefox 3.0 is fast. When it was released it was the fastest browser I had ever used. It has great javascript performance and quick page rendering. I am not an extension whore, so even today my install of Firefox is quite fast. However, based on my subjective testing Google Chrome is much faster… At everything. Pages render faster, javascript runs faster, the interface feels more responsive and faster (while at the same time feeling more slick and polished). This is a lightning fast browser. I expected decent speed, especially with javascript, but what I got was fabulous amounts of speed during every single moment of the user experience. Thank you Google! All of the problems and short comings I have found with Chrome are almost forgotten when I am constantly rewarded by the speed of this browser.

However, browser “performance” isn’t completely about speed. We must also take into account memory usage. You may recall that one of the primary goals of Firefox 3 was to not only speed up the browser, but to significantly reduce memory leaks and overall memory usage. The Firefox team met this challenge and created an extremely efficent browser when it comes to memory usage. Google Chrome does not win this battle, and I can’t say I’m surprised. In my tests Firefox uses roughly 15 – 25 MB less memory than Google Chrome, and that is with a couple extensions. Google puts up a good fight, and still beats the crap out of Internet Explorer in memory usage (who doesn’t?), but its not enough to beat Firefox. With the overhead incurred by Chrome’s process isolation and the fact that this is a beta, I can’t say I’m surprised. I am not sure if this fact will ever change, as Chrome is likely to adopt many more features before its out of beta. More features generally means more memory used, so unless Chrome can loose more weight then it adds to its tool belt it may never win the memory usage crown. We may come to find that additional memory usage is the price we must pay for the robustness and stability Chrome is aiming for. If that is the case then I will happily give up a few more megabytes.

Whats Missing Behind Google’s Shiny Chrome

I’ve spent the breadth of this article going over all of the things Google Chrome does. What I haven’t gone over is what it doesn’t do and why its not yet ready to replace Firefox or other browsers for many (or most) users.

Google Chrome does not do extensions. At all. It doesn’t call them something else or do them in some new way… as of right now Chrome just is not extensible. Also, plugin support is provided through the old Netscape API and the robustness of that support is currently unknown. However, many Firefox users (and especially extension developers) know that plugins are not extensions, and most Firefox users who make use of extensions are not yet going to make the switch to Google Chrome. When you come to rely on things like Greasemonkey, mouse gestures, Flashblock, you will not give them up easily. Hopefully Chrome will gain some form of extensibility soon. This is huge. If Chrome was extensible it would already be my default browser and I would be busy writing extensions for it.

Google Chrome is not skinnable or themeable. This is likely to change, but for now if you don’t like the default look and feel of Chrome you either have to get used to it or use another browser.

Google Chrome has a good number of other missing features and bugs. Here are some examples: There is no fullscreen mode. Chrome does not confirm when you attempt to close windows with multiple tabs… it just does it. Using the backspace key in certain rich text fields sometimes causes the browser to go back a page…

All of these are mostly minor issues, and aren’t exactly unexpected from a beta browser. They are still significant distractions though, from an otherwise great browser.

Conclusion

In the end, my final verdict is that Google Chrome is a fantastic browser that feels less like a beta than any other beta browser I have ever used. I’m not ready to give up Firefox yet, but Google Chrome has already far exceeded my expectations of what a Google browser would be like, and those expectations were already high. The paradigm shifts Google is attempting to bring to the browser world shows that they really understand where the web is heading and how browsers need to change to take us there. This is a great day for the browser market, and the fact that Google has released Chrome as an open source project makes the significance many times sweeter. I don’t imagine Chrome will be the only browser on the block to handle resources in such a robust way for long, but I don’t think Google ever intended it to be a concept exclusive to Chrome. Google’s example of what can and should be done to move the web forward will have an even greater impact on the web than the actual implementation they released today.

Just before heading off to bed last night I came across an interesting leak of information via LifeHacker,  Gizmodo, and Ars. However, the breadth of coverage wasn’t necessary to catch my attention. I knew this was big news the moment I read the headline… Google is releasing a browser, and its called Chrome. The beta is set to be released later today. Read on.

Update: Google Chrome beta has now been released. I’ll write up my first impressions soon.

There are many things about this announcement that surprised me.

First, I was suprised to learn that Google was working on a browser at all. They have a history of partnership with the Mozilla Foundation and it was always my assumption they were pretty happy with the direction of the Firefox project. Google does have a history of employing certain Mozilla Foundation members and former Netscape employees, including the former project lead for the Firefox project, Ben Goodger. Even though Ben and others have directly worked on crafting web browsers in the past, it was my honest belief that Google chose to employ them to work on things like browser compatibilty and optimization for Google’s web apps, browser plug-ins and extensions, Google Desktop tools (like Google Desktop Search), and web applications and tools for Google’s Android mobile phone platform. I also assumed these browser buffs would help build things like Google Gears and Google Web Toolkit. Who would have thought that Google’s real motivation for employing people with experience building desktop web browsers was to build a desktop web browser of their own? Okay, so maybe it should have been plainly obvious, but for whatever reason I just didn’t see this coming.

However, this is not where the surprises end…

If I had predicted Google was going to build a web browser and I had taken the time to pontificate about what that browser would look like, I don’t think my imaginary browser would have been any closer to what Google has built than, say, Firefox is to what Google has built.

For one, I would have assumed Google would start with the Gecko layout engine. Wrong. They chose Webkit, which was originally the engine for KHTML and later improved upon by Apple. Looking at this decision closer, I can’t say I would have settled on Gecko either. It appears that Webkit is much easier to program for. What I find comical is that Google named the browser Chrome, which is the name Mozilla uses to refer to user interface configurations for their various projects (Firefox, Thunderbird, Camino, Seamonkey, etc.). I am not sure if this name is a tip of Google’s hat to the Mozilla Foundation, an acknowledgement of the multiple Mozilla developers Google employs, or possibly a reference to something Google may have taken from the Firefox project: the XUL user interface language. We may find out which one, if any of these, are the reason for the name Chrome when the beta and source code are released later today

A couple assumptions I would have made about a Google browser appear to be true: the interface is simple, and they have attempted to improve javascript performance and robustness. Given the company’s much lauded approach to very simple user interfaces, and their reliance on using javascript as an application framework, I am not surprised by either of these features. Their V8 javascript rendering engine is claimed to be very fast and robust. I will have to wait for the beta to see just how true this is.

The next big difference between Google Chrome and other browsers is the way it handles resources. Every single tab or window is a seperate operating system process, with isolated memory and rendering resources. Leaked screenshots even show that the browser will include its own internal task manager, should you need to swat a tab out of existence. What is the significance of this? Well, the idea is that should some plugin or website crash your browser, it won’t crash your browser… it will only cause that website’s tab or window to close, leaving the remaining tabs/windows unaffected. This also means that when you close a tab you are killing a process, so every last byte of memory used by that tab is reclaimed, helping thwart nasty memory leaks. I feel that this is Chrome’s “killer feature”. Google wants web applications to replace desktop applications, and for that to happen a browser must be more than just a browser. It must be a platform. It must be an operating system for web applications. This means it must have the robust process isolation and resource management you expect from an operating system. This idea is so simple yet so powerful, I am shocked it has not been done [right] before. I’ll still have to wait for the beta to decide just how well Google Chrome does this.

This is a significant day for web developers, application developers, and the internet at large. I am much more excited than I thought I would be about the concept of a Google Browser. Welcome, Google Chrome!