backtop


Print 57 comment(s) - last by dramatic.. on Sep 10 at 4:47 PM

What browser is the best? We have the information you need to decide

Last week Opera 10.0 was released in its final form.  DailyTechhad covered the development process of the browser exhaustively, so some may have been surprised not to see a piece on the browser.  That's because we were working on something, special -- a complete review of not just Opera 10.0, but all next generation browsers.

We have extensively benchmarked all of the next generation browsers -- Safari 4, Opera 10.0, Firefox 3.6 alpha 1, Internet Explorer 8, and Google Chrome 3 and 4.  In this first segment we will look at basic features, install time, and browser launch times of these next gen browsers, in comparison to previous editions.  In the second segment, we will look at CPU and memory usage, and briefly look at browser security.  And in the third and final segment, we'll look at performance in CSS and Javascript benchmarks, as well as a basic rendering benchmark.

1.  User Interface and Basic Features

Let's first look at browser features:

Browser License Prompts to Make
Default

on Install
Prompts to Make Default on Open Uninstall Shortcut Uninstall Option to Delete Personal Data Spell Checking Colorized Tabs Favorites Tiled Homepage Built In Mail Client Compression Boost
Option
Opera 9.6 Proprietary No Yes, popup No No Yes No Yes Yes Yes
Opera 10.0 (same) Yes, Unchecked Yes, popup No No Yes No Yes Yes Yes
Firefox 3.5 MPL, MPL/GPL/
LGPL tri-license, Mozilla EUL
Yes, Unchecked Yes, popup Wizard Via Control Panel Yes Yes Via add-on only Via add-on only Via add-on only No
Firefox 3.6 (same) Yes, Unchecked Yes, popup Wizard Via Control Panel Yes Yes Via add-on only Via add-on only Via add-on only No
Chrome 2 source-
BSD Executable – Google Terms of Service
Yes, Prechecked Yes, in frame Yes Yes Yes No Yes No No
Chrome 3 (same) Yes, Prechecked Yes, in frame Yes Yes Yes No Yes No No
Chrome 4 (same) Yes, Prechecked Yes, in frame Yes Yes Yes No Yes No No
IE 8 Proprietary Yes, Unchecked Yes, popup No, and harder to reach in Control Panel No Add-On (IE-Spell) only Yes No No No
Safari 3 Engine – GNU LGPL Everything Else - Proprietary No Yes, popup No No Yes No No No No
Safari 4 (same) No Yes, popup No No Yes No Yes No No


As you can see, when it comes to the user interface, Opera arguably leads the pack, with the most built-in interface features (favorites tiled homepage, built in mail client, server-side compression, etc.).  Firefox is a close second, with the most features supported, if you install add-ons (though add-ons decrease performance and can cause compatibility issues).  This assessment does not include security, which we'll look at in a later piece.

2. Installation

Google's Chrome is the easiest to install.  Microsoft's Internet Explorer 8, on the other hand, is the hardest to install, by far, requiring a full system reset. In retrospect, the controversy over IE 8 asking at installation if you want to make it your default browser, seems rather silly.  In fact, all the browsers ask you this on startup, and Google even goes so far as to pre-check an option during the install to make Chrome the default browser. We do like how Google's post-install browser check is incorporate as a less obtrusive browser window frame, rather than the popup that the others used (all the browser allow users to permanently dismiss these inquiries).


In our first benchmark, we did a clean install of each browser and measured the time the install took, using a lightweight, hotkey-driven timer for time measurement.  As you can see from the chart to the right, Internet Explorer (as with the uninstall) was the worst, taking the most time (by-far) and requiring a full system restart. Chrome 4 was the best, taking a mere 11.4 seconds to install after double clicking the installer package (Note: Time to import bookmarks, etc. was not included in this time, just the time to complete the actual install).

3.  Application Launch Speed

Our second benchmark looks at browser launch times.  We again used the hotkey timer and this time measured the time it took for the browser window to appear after double clicking.  Page-load was not necessary, the times measured indicated the time to get the address bar to a responsive state.


In our first run, we launched the browser "cold" after a full system restart.  Averaged over three trials, Chrome by-far launched the fastest, with Chrome 4 being the fastest of the Google browsers.  In close second was Opera 10.0.  Firefox 3.5 was strangely slow, taking 10+ seconds in more than one trial, so we launched it five times (stock install, no add-ons).  We finally concluded, that this was expected and not an error.  In the end Firefox 3.5 was the slowest to launch, but Namaroka (Firefox 3.6 alpha 1) launches much faster.  IE 8 was the second slowest in the cold launch trial.


In our second run we launched the browser "warm" after having already launched it once.  As can be expected, due to caching, the browsers launched significantly faster this way.  For warm starts, Chrome was yet again the fastest, while Opera and Safari came in second, though both of these browsers curiously showed slower start times in their latest versions (Opera 10.0 and Safari 4) than in their previous versions (Opera 9.6 and Safari 3).

In the next segment we'll look at memory and CPU usage.  We'll also examine how the browsers stack up in security.

Note: Note: All benchmarks were performed in 32-bit Vista on a Sony VAIO laptop with 3 GB of RAM, a T8100 Intel Processor (2.1 GHz), and a NVIDIA 8400 GT mobile graphics chip. The number of processes was kept consistent and at a minimum to reflect stock performance.


Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Acid3 & HTML5 Comparison
By forever4now on 9/8/2009 9:08:32 AM , Rating: 2
One topic that is often overlooked, when comparing browsers, is conformance to web standards.

It would be great, if you could provide a summary of Acid3 test results, for each browser. The Acid3 test is here:

http://acid3.acidtests.org/

It would also be useful, if you could provide a summary of key HTML5 functionality supported by each browser.

For example, try viewing the following HTML5 video, with IE 8 (or earlier). You CANNOT!

http://demo.sproutcore.com/video/

Try viewing this same video with Chrome or Safari. You CAN! And, it looks AWESOME!




RE: Acid3 & HTML5 Comparison
By JasonMick (blog) on 9/8/2009 9:32:08 AM , Rating: 3
Hi,
In the (upcoming) third segment I include some information on standards not supported in each of the browsers.

Acid 3 results will be included.

Also the Peacekeeper synthetic benchmark performance will be discussed -- Peacekeeper looks at CSS, Javascript, rendering and more, and that score also reflects the level of standards support.

Stay tuned for the next two sections, but feel free to let me know if you think there's something missing/something you want to see!

-Jason


RE: Acid3 & HTML5 Comparison
By Cypherdude1 on 9/9/2009 6:51:42 PM , Rating: 2
quote:
Firefox 3.5 was strangely slow, taking 10+ seconds in more than one trial, so we launched it five times (stock install, no add-ons). We finally concluded, that this was expected and not an error.

My "cold" starts for FireFox 3.5.2 is faster than what you report. "Cold" starts from a fresh boot is 7 seconds, not 10. "Warm" starts for me are in line with what you report, 3.0 secs. Aside from a number of small bugs, FF 3.5.2 gives satisfactory performance, no crashes or blue screens.

The only TSR's I have in RAM from boot are the bloatware Norton Anti-Virus 2006 (from SystemWorks 2006 with all other TSR features disabled), ATI Radeon Taskbar Tray icon, and 3 other smaller unrelated RAM TSR Tray icons.

The interesting part is that I have an 8 year old system running WinXP SP3: AMD 1400 T-Bird (single core), 1 GB DDR1 PC2100 RAM, ATI AGP 8x (running @4x because of mobo) Radeon video card, PCI Hercules sound card. The only part of my system which is not (much) slower than modern systems is my hard drive: Seagate 7200 RPM Ultra100 250 GB's PATA HDD. I have tweaked WinXP SP3 as much as possible. While I don't play games on it anymore, it runs fine for all other apps, even playing video.

I say bloatware Norton SystemWorks 2006 (many bugs included for free), because before I installed SW2006 my boot time was 90 secs. Now it's 140 secs , a 50 second bonus. All these extra features are given by Symantec to their customers because they appreciate our business so much. B^D


RE: Acid3 & HTML5 Comparison
By Fanon on 9/8/2009 10:30:51 AM , Rating: 3
Acid is the 3DMark of the web. Is it really the all important test its made out to be? No.

And what's so important about HTML5 features? It's not even close to being a final spec. Hell, they just changed the footer element. Any HTML5 feature added by any browser is fair game to be changed, and then you'll cry foul when a browser's implementation differs from the non-existent spec.


RE: Acid3 & HTML5 Comparison
By forever4now on 9/8/2009 11:34:51 AM , Rating: 1
Acid3 tests standards conformance so YES, that DOES make it important. You can read about Acid3 in Wikipedia.

Re. HTML5: Most of the major browsers have already rolled out HTML5 features, so it would be worthwhile to have a table to see who supports what. Many of the features will probably reach de facto standard status, before the spec is fully approved.


RE: Acid3 & HTML5 Comparison
By Fanon on 9/8/2009 12:17:02 PM , Rating: 3
I know what Acid tests. The problem with Acid is it only tests certain selected features/spec support. It's 3DMark for web browsers. Whooop dee dooo.

Indeed, most browsers have implemented some HTML5 features, but my point still stands. HTML5 is only important to a select group of people. That group will undoubtedly grow as HTML5 nears completion in 4497 days, but the "spec" isn't even close to being finished, and fundamental aspects change frequently. It's far too early to clamor for HTML5 features yet.


RE: Acid3 & HTML5 Comparison
By forever4now on 9/8/2009 1:06:09 PM , Rating: 2
Standards conformance is a VERY important aspect of browsers. It affects how web developers build websites. Acid3 provides a measure of how well each browser conforms, so it is VERY appropriate to include, with a browser comparison.

Re. HTML5: I think it will be rolling out more quickly & more broadly than you are imagining, but that is not the point of the table. It's just to list the HTML5 features supported by each browser. Simple.

If you, personally, don't have an interest in these things, then you should ignore them, when they are published.


RE: Acid3 & HTML5 Comparison
By Fanon on 9/8/2009 3:18:28 PM , Rating: 2
I have an interest in these things; I have for 10+ years.

I never said, nor implied, standards aren't important. I criticized the importance of Acid3. Two completely different topics.

HTML5 isn't important or relevant today. The spec is always changing, and the worth of the features implemented is debatable as only the latest browsers implement them. As the spec changes, and thus changes to the implementation, you're going to end up with a fractured web. Implementing specs when they're barely in an embryonic stage hurts developers in the long run.

I think you're too optimistic in HTML5's release. Hell, even the author says no final spec until 2020-something.


RE: Acid3 & HTML5 Comparison
By forever4now on 9/8/2009 4:25:57 PM , Rating: 2
Considering all the effort Mozilla, Google, Apple & Opera put in to passing the Acid3 test, I would say it is probably important.

Considering that some of the major browsers are already shipping with HTML5 functionality built in, I would say it is probably important.

I agree that the entire HTML5 spec (which is quite complex) may take some time to ratify, but the fact is, many HTML5 features are already in production browsers (which is why I suspect that many HTML5 features will become de facto standards, before the spec is even ratified).

Anyway, what's the big deal?

If you search Wikipedia for "Acid3" and "HTML5", you could almost derive these tables from the information listed there. I'm just suggesting that a couple of simple summary tables be included, to make the browser comparison article complete.


RE: Acid3 & HTML5 Comparison
By neilrieck on 9/10/2009 7:01:12 AM , Rating: 2
IIRC, HTML5 is not much more than strict XHTML which can be enabled in today's compliant browsers just by inserting the correct DOCTYPE.


RE: Acid3 & HTML5 Comparison
By adiposity on 9/8/2009 2:56:14 PM , Rating: 2
Works with firefox...


RE: Acid3 & HTML5 Comparison
By Justin Time on 9/9/2009 10:31:10 AM , Rating: 2
>> One topic that is often overlooked, when comparing browsers, is conformance to web standards.

Standards are a good thing, but you can't have it both ways and say that standards includes non-standards.

>> It would be great, if you could provide a summary of Acid3 test results, for each browser.

One of the problem with Acid3 is that it tests for CSS3 properties, and CSS3 is not a w3c standard.

>> For example, try viewing the following HTML5 video, with IE 8 (or earlier). You CANNOT!

Again, html5 is NOT a web standard, it's still a working draft and subject to change.

MS was demonised in the past for implementing non-standard features, and was burned by implemented features that were in working draft status, only to have them pulled from the candidate standard.

They now seem to be taking the conservative approach that they will not implement features until they reach candidate status, and are being accused of dragging their feet by not implementing non-standard features.


RE: Acid3 & HTML5 Comparison
By forever4now on 9/10/2009 2:24:49 AM , Rating: 2
"CSS3 is not a w3c standard"
"html5 is NOT a web standard"

Isn't it a little strange that everyone, with the exception of Microsoft (i.e. Mozilla, Google, Apple & Opera) is deploying browsers that support this stuff?

Working implementations are VERY important to the standardization process. They help to validate/solidify the specifications. Tests, like Acid3, help to ensure the implementations are consistent.

Microsoft needs to stop making excuses and JOIN the international standards community, instead of FIGHTING it.


RE: Acid3 & HTML5 Comparison
By neilrieck on 9/10/2009 6:54:21 AM , Rating: 2
You are correct. When Microsoft defeated Netscape during the browser wars in the mid 1990s, they were only thinking about themselves. Then they didn't do very much with IE for the next 10 years.

Can you imagine the cross-browser testing that must go on each day at big sites like eBay and Amazon? Since the appearance of standards-compliant browsers, "test-fix" usually means adding special hooks to handle every version of IE. What a mess.


"We’re Apple. We don’t wear suits. We don’t even own suits." -- Apple CEO Steve Jobs

Related Articles













botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki