Linked by Thom Holwerda on Tue 23rd Mar 2010 23:57 UTC
Internet Explorer "At last year's PDC, held in November, Microsoft showed a graph showing scores of a variety of Web browsers in the SunSpider JavaScript benchmark, to show off the progress that the company was making with Internet Explorer 9. Another such graph was shown off at the recent MIX event. What was most interesting about the graph was not IE9's progress, but Opera's."
Thread beginning with comment 415054
To read all comments associated with this story, please click here.
Diminishing returns
by nt_jerkface on Wed 24th Mar 2010 08:36 UTC
nt_jerkface
Member since:
2009-08-26

The focus on synthetic javascript benchmarks really needs to be questioned. When web surfing the bottleneck is your bandwidth and even after that the typical website does not make heavy use of javascript.

There's also the question of whether or not users can even tell the difference past a certain threshold.

Reply Score: 2

RE: Diminishing returns
by wirespot on Wed 24th Mar 2010 08:46 in reply to "Diminishing returns"
wirespot Member since:
2006-06-21

When web surfing the bottleneck is your bandwidth


Is it really? We seem to get more and more bandwidth that I dunno what to do with. In Europe, at least, speeds of up to 20 Mbps seem to be common, and even the lowest are still a few Mbps.

and even after that the typical website does not make heavy use of javascript.


Doesn't it? I develop a large community website and I find myself adding JavaScript all the time, for any number of purposes, from pretty graphical effects to AJAX. And this is just the most common stuff, I'm not even talking about a site that actually tries to do really fancy things.

There's also the question of whether or not users can even tell the difference past a certain threshold.


Have no worry, whenever more [computing] power is available, people usually quickly manage to max it out. ;)

Edited 2010-03-24 08:47 UTC

Reply Parent Score: 6

RE[2]: Diminishing returns
by nt_jerkface on Wed 24th Mar 2010 09:11 in reply to "RE: Diminishing returns"
nt_jerkface Member since:
2009-08-26


Is it really? We seem to get more and more bandwidth that I dunno what to do with. In Europe, at least, speeds of up to 20 Mbps seem to be common, and even the lowest are still a few Mbps.

That's peak speed and it says nothing about the server sending you the file. Network latency is more of an issue than javascript processing time.


Doesn't it? I develop a large community website and I find myself adding JavaScript all the time, for any number of purposes, from pretty graphical effects to AJAX.

The typical website is mostly HTML, not Javascript. Even a lot of the AJAX heavy websites still generate a lot of HTML since most of the processing is done server side.


Have no worry, whenever more [computing] power is available, people usually quickly manage to max it out. ;)

It has nothing to do with computing power.

Once people can no longer tell the difference between super-fast and ultra-fast rendering the question needs to be asked if these benchmarks even matter. I think to often they are just used for cheap bragging rights.

Reply Parent Score: 5

RE: Diminishing returns
by lemur2 on Wed 24th Mar 2010 09:20 in reply to "Diminishing returns"
lemur2 Member since:
2007-02-17

The focus on synthetic javascript benchmarks really needs to be questioned. When web surfing the bottleneck is your bandwidth and even after that the typical website does not make heavy use of javascript.

There's also the question of whether or not users can even tell the difference past a certain threshold.


In the article linked here:
http://arstechnica.com/open-source/news/2009/02/mozilla-demos-impre...

there is a video of a demo run by Christopher Blizzard at SoCal Linux Expo which shows Javascript Motion Tracking.

You can watch the demo here:
http://www.youtube.com/watch?v=W0Hl0tLs7XA&feature=player_embedded

The motion tracking software is running, as Javascript, within Firefox, which is also playing the video.

Users certainly could tell the difference if anyone tried to do this type of thing running Javascript at IE speeds.

Javascript is also the means by which one animates SVG graphics to achieve (in a standards compliant way) the equivalent of effects like Flash (swf) or Silverlight animations. If IE9 keeps the same dog-slow javascript performance as IE8, then Flash or Silverlight plugins would still be necessary to achieve even partially acceptable animated graphics effects.

http://en.wikipedia.org/wiki/SVG_animation
http://srufaculty.sru.edu/david.dailey/svg/SVGAnimations.htm

One definitely needs the very best javascript performance one can get in order to have a good experience when viewing the standards-compliant web.

This is the entire reason why standards-compliant browsers pay such attention to dynamic javascript performance.

Does this soccer ball spin for you (can you even see the soccer ball)?

http://upload.wikimedia.org/wikipedia/commons/4/4f/Soccer_ball_anim...

If it doesn't spin, or you can't see it, then you haven't got standards-compliant animated graphics.

Once people can no longer tell the difference between super-fast and ultra-fast rendering the question needs to be asked if these benchmarks even matter. I think to often they are just used for cheap bragging rights.


Nope. They are used for graphics animations, and rich interfaces, instead of proprietary rubbish like Flash or Silverlight.

Like this:

http://tavmjong.free.fr/INKSCAPE/DRAWINGS/clock2.svg

Edited 2010-03-24 09:33 UTC

Reply Parent Score: 5

RE[2]: Diminishing returns
by nt_jerkface on Wed 24th Mar 2010 09:44 in reply to "RE: Diminishing returns"
nt_jerkface Member since:
2009-08-26


The motion tracking software is running, as Javascript, within Firefox, which is also playing the video.

Users certainly could tell the difference if anyone tried to do this type of thing running Javascript at IE speeds.

That doesn't represent typical browsing but more importantly my point had nothing to do with IE or any specific browser. It's a question of diminishing returns much like an improvement to a 3D gaming engine that already displayed a consistent frame rate well beyond what the eye can perceive.


...then Flash or Silverlight plugins would still be necessary to achieve even partially acceptable animated graphics effects.

Those plug-ins will be around for at least another 5 years so don't expect a javascript revolution.

Reply Parent Score: 2

RE[2]: Diminishing returns
by flanque on Wed 24th Mar 2010 09:45 in reply to "RE: Diminishing returns"
flanque Member since:
2005-12-15

http://upload.wikimedia.org/wikipedia/commons/4/4f/Soccer_ball_anim...

Using Firefox 3.6.2 and it shows but doesn't spin. I thought FF was standards compliant?

Reply Parent Score: 2

RE[2]: Diminishing returns
by smashIt on Wed 24th Mar 2010 10:44 in reply to "RE: Diminishing returns"
smashIt Member since:
2005-07-06

Javascript is also the means by which one animates SVG graphics to achieve (in a standards compliant way) the equivalent of effects like Flash (swf) or Silverlight animations. If IE9 keeps the same dog-slow javascript performance as IE8, then Flash or Silverlight plugins would still be necessary to achieve even partially acceptable animated graphics effects.


if IEs bad js-performance is what stops webdesigners from using that crap than we should thank ms for it

Reply Parent Score: 2

RE: Diminishing returns
by lemur2 on Wed 24th Mar 2010 09:51 in reply to "Diminishing returns"
lemur2 Member since:
2007-02-17

the typical website does not make heavy use of javascript.

There's also the question of whether or not users can even tell the difference past a certain threshold.


http://en.wikipedia.org/wiki/SVG_animation

Scripting: ECMAScript is a primary means of creating animations and interactive user interfaces within SVG.


Not only graphics animations, but also interactive user interfaces.

One certainly doesn't want lack-luster/sluggish performance when using an interactive user interface.

Reply Parent Score: 3

RE[2]: Diminishing returns
by nt_jerkface on Wed 24th Mar 2010 10:28 in reply to "RE: Diminishing returns"
nt_jerkface Member since:
2009-08-26


Not only graphics animations, but also interactive user interfaces.

One certainly doesn't want lack-luster/sluggish performance when using an interactive user interface.


At some point those user interfaces feel equally fast even if one browser is actually rendering them slightly faster. That is the point I was making, but I can see you just tried to turn my point into your own HTML5 advocacy piece.

I think benchmarks should be related real world use, especially if they are being used to rate software. I think it is a joke when people make an issue of Opera or Chrome beating the other in a synthetic Javascript benchmark when for the heaviest AJAX websites it is doubtful that anyone can tell the difference.

It's called diminishing returns and I predict this will become even more apparent with the next generation of browsers. But I'm sure we'll have the same synthetic benchmarks and a "browser speed king" even if the differences are imperceptible.

Reply Parent Score: 2

RE: Diminishing returns
by kaiwai on Wed 24th Mar 2010 13:30 in reply to "Diminishing returns"
kaiwai Member since:
2005-07-06

The focus on synthetic javascript benchmarks really needs to be questioned. When web surfing the bottleneck is your bandwidth and even after that the typical website does not make heavy use of javascript.

There's also the question of whether or not users can even tell the difference past a certain threshold.


It is the chicken and egg scenario of developers could use more javascript if they wanted to but at the same time given the number of crappy and slow performing browsers out there, they have to take that into account when designing their site. So what you have are websites that could load faster because of smarter ways of updating a given web page but instead are bound to using less javascript - gmail where the whole page doesn't need updating when a new email arrives as an example.

Reply Parent Score: 2

RE[2]: Diminishing returns
by nt_jerkface on Wed 24th Mar 2010 19:39 in reply to "RE: Diminishing returns"
nt_jerkface Member since:
2009-08-26


It is the chicken and egg scenario of developers could use more javascript if they wanted to but at the same time given the number of crappy and slow performing browsers out there, they have to take that into account when designing their site.


Being held back by IE6 especially has long been a problem for web developers but that still says nothing about how much of a return there is between the latest improvements in Chrome versus Firefox.

It's become like a framerate benchmark where the perceptible difference is no longer visible to the user. It would be like rating one video card over another because it can render a synthetic benchmark slightly faster, even though both are rendering past 100fps.

But then I suppose that the tech press is filled with browser fanboys (cough cough ZDNet) that will take any test as long as it shows their favored browser performing x% better. Who cares if the difference is imperceptible, it's x% better and that's a headline!

Reply Parent Score: 2

RE: Diminishing returns
by arpan on Wed 24th Mar 2010 20:04 in reply to "Diminishing returns"
arpan Member since:
2006-07-30

The web no longer works only on the desktop. These browsers are going to be used in Mobile devices and JS performance will make a noticeable difference in usability and battery performance.

I'm sure one main reason why Apple, Google & MS are focusing so much on JS performance is their mobile OS. It can't be a coincidence that MS has started concentrating on IE and plans to release it at the same time that they have made a massive change to their phone OS and plans to release it. If Windows Phone 7 is released with IE8, the web experience is not going to be very good in comparison to the iPhone & Android devices.

Reply Parent Score: 2

RE: Diminishing returns
by cerbie on Wed 24th Mar 2010 23:35 in reply to "Diminishing returns"
cerbie Member since:
2006-01-02

The focus on synthetic javascript benchmarks really needs to be questioned.
That I can agree with.

When web surfing the bottleneck is your bandwidth
Generally not so much, with ABP and Noscript keeping out all those 3rd-party loaded files.

and even after that the typical website does not make heavy use of javascript.
Sadly, that is both true and false. In terms of features, it is true. However, it is rather false, as javascript often gets used to do things that HTML has been doing for many years, and performs overly-complicated tasks for very simple things, because it's just so awesome what you can do with jquery, once you are loading 50k worth of code...

There's also the question of whether or not users can even tell the difference past a certain threshold.
Jsmath can cause browsers to temporarily freeze on C2Ds, at the least. I'm also pretty sure the "proper" DOM mods (instead of innerHTML) are slow and CPU-hogging, except where they are done often enough to benefit from binary compiling and other optimization methods.

That said, the real bug is that javascript engines seem to, across browsers, be able to negatively affect the responsiveness of the web browser. if that were taken care of, the speed of javascript itself, while still important, would be secondary. It would especially be nice, since more typical computing power leads to more of it being used (I am of the opinion that software should be lean and responsive, to give users more freedom to squander resources without worry). IE 7 and 8 have been lagging very far behind practically all other browsers in that regard. It's a problem for any I've used, but IE8 makes me think I'm back on my Cyrix PR150+ from all the freezes while it loads all the JS, or has to do lots of DOM work...

Edited 2010-03-24 23:41 UTC

Reply Parent Score: 2