Browser VS Browser - How they scored

[nkuvu]

Quote: Originally Posted by roncross@cox.net All of these browsers should be scriptable with respect to launching urls. A simple code would look something like the following

That was part of my intent, but not all of it.

For example, what I was thinking of was a script that could do the actual timing, clear the cache, et cetera. I'm not sure that AppleScript could do the job, but I'm also not too familiar with AppleScript capabilities.

[macmath]

Quote: Originally Posted by Xd To be honest with you, I have not noticed that yet(fingers crossed). This is why we hoped someone else would do the tests also. I like the look and clean functionality of safari. Explorer was so unresponsive that anything would have been better. I am not a promoter of camino or a basher of other browsers, I really just thought it would be interesting to see which free version works fastest but it should be noted that we are after all talking about mere seconds for the casual user. If you are online all day those seconds turn into 5-10 minutes of "waiting" time and by the end of the day/week it becomes really significant.

I agree that what you addressed is important and found the thread to be interesting and of value. Do you have a significantly faster machine? Maybe there are some settings which I could turn off or on the remedy the slowness on my machine. Thanks.

[Xd]

This seems achievable although I have no experience writing scripts. I just think it is important that someone other than me do a similar test so that we can get a comparison. My results should be challenged. Any other results are welcome here.

[mervTormel]

Xd, could you explain how to repeat your tests? how did you time to hundredths of a second? what methods did you use to setup the tests? what other processes were running on your rig? how were the browsers configured?

regrets if this was explained earlier in the thread; i can't find it.

[Xd]

Hey macmath the possibilities could be endless. To start off with, I have a G3 400 mHz, 320mb ram. A really humble machine. Just a straight install of panther 10.3 a week ago, give or take a day, no tweaks. It may come down to your connect speed. I have a download speed of 2871 kbps and an upload speed of 623 kbps. For comparison value speed really makes no difference, it comes down to consistency. In other words the place of finish for each browser.

[Xd]

Hey mervtormel any timer will do. Mine goes to the thousandth second and I rounded up. I wrote the order of sites to visit just as I indicated in the first post and followed the order and recorded the times. No duplicate tests were ever done for fairness in cache. I cleared the cache at the beginning for each browser and went through them one by one so there was the smallest amount of time between browsers. No tweaks to any of the browsers, just stock. We all know how different times of the day cause faster and slower loads so I wasted no time. I waited 12 hours and repeated the tests for the 2nd study. The 3rd test took the longest because I randomized the order of browsers and also the sites but, they were done at the same time one after the other. The sites were the same only the order was shuffled around. I posted the results. Nothing running while I performed the experiment for consistency.

[bramley]

Yes, Applescript could be used to time page loads for browsers that supported Javascript commands like "document.readyState" to find out if the page has been loaded.

Thinking about this a bit more, I came to the conclusion that different browsers probably implement their Applescript in different ways. So by relying on Applescript to do the donkey work, you would just be introducing another variable into the mix. You can compensate for it. You'd need to test each browser manually first and compare it to when it was working on "automatic." And the best way to do that must be to load a webpage off your hard drive, and compare times.

Which sparked another thought. What is the value of loading a web page from the web? As I think was pointed out previously, the speed of loading is strongly influenced by your connection, other users, etc. - factors which are unaffected by your choice of browser. If you loaded all tests from the hard drive from a previously agreed set of downloaded local pages, you eliminate all of these variables. You do introduce your hard drive into the equation, but the results would be at least be consistent per machine, and not influenced by external factors at all.

[Xd]

If someone can walk me throught the set-up steps I will redo the test with an autoscript. The less human interference the better. I would need to know each specific step as I have little experience to go on.

[mervTormel]

Quote: Originally Posted by bramley ...Which sparked another thought. What is the value of loading a web page from the web? As I think was pointed out previously, the speed of loading is strongly influenced by your connection, other users, etc. - factors which are unaffected by your choice of browser. If you loaded all tests from the hard drive from a previously agreed set of downloaded local pages, you eliminate all of these variables. You do introduce your hard drive into the equation, but the results would be at least be consistent per machine, and not influenced by external factors at all.

this was hinted at earlier and further exhorts an answer to the question, "what are we testing?" and "are the methods sound?"

Xd's "timer" is still a mystery to me.

[Xd]

My timer is a little invention that came out several decades ago called a stopwatch. I press the little button simultaneously at the launch of a web page and press stop when the page is loaded. Then I record the times. I am really trying to see which free browser on average performs the fastest. The methods are if nothing else consistant for each load. Oops, the orig sounded sarcastic, not meant.

[mervTormel]

how do you "launch" a web page? how is your reaction time? do you drink coffee or beer? what constitutes that the page load is considered complete? are there any load errors? what are the contents of your /etc/hosts file? what other network/competing processes are running?

not meant to be snarky. just that the parameters of the tests are vague. granted, they give a feel of performance, but i can get that without timings to two decimal places.

it would be interesting to validate performance metric comparisons of others, as long as we know the environment is consistent. otherwise, statistics gathered from erroneous data are misleading.

[Xd]

I am not offended at all. It is no intention of mine to proclaim I am the definitive browser tester. In fact, I was hoping to inspire others to challenge my times and repeat the tests to compare results.

I will use safari as an example for how I do each test in hopes of clearing up any misconceptions.

I Open safari or any browser, clean out the cache. In the address bar I write in a web-site, like www.espn.com. Next I set the timer to zero and simultaneously press start of the timer/Return on the keyboard. When the load bar is finished I press stop on the timer. Omniweb however uses the grey clockwise load circle on the taskbar. The rest use the standard blue highlight bar which turns transparent upon completion of page. I stop timing upon completion and record the time.

My reaction time/consistency is good but even if I state that, it should still be challenged which I hope someone takes up. About 25 posts back we discussed automation but nobody came back with specific enough clarity for me to try that approach.

[blubbernaut]

I'm lovin' this thread, and in fact have started using Camino since reading through this. I do like it's feel, but I really think someone needs to write a variety of different sites using the aforementioned Javascript:

Quote: Originally Posted by bramley Yes, Applescript could be used to time page loads for browsers that supported Javascript commands like "document.readyState" to find out if the page has been loaded.

Then some automation, perhaps not on locally hosted sites, but perhaps on some kind soul's remote host that is password locked to prevent speed fluctuations due to the rest of the world accessing the same server! A big ask, I know, but while we're are wishing...

This guy (who has worked on both Safari and Mozilla/Camino) has confused me regarding Camino's speed... can anyone else translate?

[bramley]

Well, as it turns put, this is harder than I'd thought. The script below is for timing Safari, and works fine. I tried to do the same for Camino - and for some reason I cannot open the AS dictionary for Camino - so I can't get the syntax right. The other problem is that the AS 'clock' only has a resolution of 1 second, so although my script is theoretically accurate to 0.1 s, I am unable to display this.

I'm a little pressed for time, so I'm posting the script and hopefully somebody can work on the variants needed for the different browsers. Also if someone has a bit of PERL/UNIX skill they could sort out a timer with better resolution.

Here is the script:

Click to open this script in Script Editor 
tell application "Safari"
	activate
	set response to display dialog "Empty cache first?" buttons {"No", "Yes"}
	if button returned of response = "Yes" then
		tell application "System Events"
			tell process "Safari"
				keystroke "e" using {command down, option down}
				keystroke return
			end tell
		end tell
		delay 5 --pause while cache is emptied
	end if
	set response to display dialog "Enter target URL" default answer "http://www.macosxhints.com" --Advertising only! DON'T actually test on this site!
	set the_URL_to_open to text returned of response
	set start_time to current date
	set the_doc to make new document at the end of documents
	tell the_doc to set URL to the_URL_to_open
	delay 1
	repeat
		if (do JavaScript "document.readyState" in front document) is "complete" then
			set stop_time to current date
			exit repeat
		else
			delay 0.1
		end if
	end repeat
	set total_time to stop_time - start_time
	display dialog "Time taken [s] to load the page was: " & total_time as string
end tell

[roncross@cox.net]

I tried opening up Camino dictionary also and it will not open. I don't believe they have one. I sent a message to them asking them how to open it up. I will post back when I hear from them.

thx
RLC

[macmath]

Quote: Originally Posted by roncross@cox.net I tried opening up Camino dictionary also and it will not open. I don't believe they have one. I sent a message to them asking them how to open it up. I will post back when I hear from them. thx RLC

I had the same problem with the FireFox dictionary some time back (seethis thread). I believe that it has one but that there is a bug which prevents its opening. I think that every application has one with a standard (minimal) suite that Apple requires.

[roncross@cox.net]

Camino's reply to my email inquiry:

this is covered in the release notes:

http://www.mozilla.org/products/cami...eases/0.8.html

"If you try to open Camino AppleScript dictionary, say from Script
Editor, you’ll get an error. To work around this, open the application
package, and remove the “Localized.rsrc” file in
Contents/Resources/English.lproj. This will, however, prevent Shockwave
Director content from loading so we recommend you do this on a copy of
the application."

I made a copy of Camino as requested and afterward I was able to see Camino script editor dictionary. It support JavaScript which is something that I am fond of using.

I put the same request in for firefox. They don't seem to have a localized.rsrc file in the English.lproj directory. I will post when I hear more.

Thanks
RLC

[Xd]

Yeah these are some of the seemingly endless variables in trying to conduct a fair speed test for these browsers. Few would dispute that safari completely dominates Internet Explorer you can actually SEE it with your eyes. Its like we all know automation would eliminate any hint of discrepency but the same script doesnt apply to each browser. Once we have that, then we will have the tools for definitive results. This would be the proof of seal for our efforts that would do more to validate the study than to prove varying results.

[Xd]

As if explorer needs more bad news.

read:

http://story.news.yahoo.com/news?tmp...5524_2004jun25

[Xd]

Bad link above, I can not edit it and appearantly those links have a shelf life. Basically it was an article concerning IE's security flaws which Msoft is now claiming to have addressed.