Otherwise, I also tested on Firefox and Safari on OS X where there are (I think) more accurate timers.
-- dz On Thu, Apr 2, 2009 at 12:01 PM, David Zhou <[email protected]> wrote: > I booted into two dedicated Windows partitions that have IE6 and IE7 > installed respectively. This was on a Core 2 Duo 2.4, so YMMV. > > -- dz > > > On Thu, Apr 2, 2009 at 11:25 AM, Andrea Giammarchi > <[email protected]> wrote: >> Those times are averages and I performed the test 5 times for each browser. >> >> I would like to know if you used IE tester ( my choice ) rather than >> Multiple IE >> >> On Thu, Apr 2, 2009 at 4:17 PM, David Zhou <[email protected]> wrote: >>> >>> I did, and it doesn't quite match up to your times -- which could be >>> due to machine differences or any number of reasons. Pure DOM was >>> still significantly faster, but not as much as the numbers shown in >>> your case. >>> >>> This is why I asked for more details about the tests you ran. >>> >>> -- dz >>> >>> >>> >>> On Thu, Apr 2, 2009 at 10:41 AM, Andrea Giammarchi >>> <[email protected]> wrote: >>> > if you want to try the test you can visit this page: >>> > http://bill.dojotoolkit.org/taskspeed/ >>> > >>> > or create a local TaskSpeed implementation >>> > >>> > On Thu, Apr 2, 2009 at 3:19 PM, David Zhou <[email protected]> wrote: >>> >> >>> >> How many times did you run your tests? Are these averages? >>> >> >>> >> Also, did you try the tests on a non-Windows XP or Vista machine to >>> >> get more accurate results? (At least for Firefox and Opera). >>> >> >>> >> I don't doubt that pure DOM manipulation is going to be faster, but >>> >> it's certainly surprising to see how much faster it is than any >>> >> library. >>> >> >>> >> -- dz >>> >> >>> >> >>> >> >>> >> On Thu, Apr 2, 2009 at 10:03 AM, Gilles <[email protected]> wrote: >>> >> > >>> >> > I understand better your point, but I think to an extend people use >>> >> > libraries for left side of the equation like you said, which is >>> >> > probably why library developers seems to concentrate more on that. >>> >> > Also I think JQuery is on the right track as the latest releases were >>> >> > about improving performance rather than adding some other cool stuff >>> >> > to it. Frankly I think it got enough cool stuff already, but I am >>> >> > looking forward to the require() addition tho. >>> >> > >>> >> > The thing is tho, if a library doesn't support the cool stuff it will >>> >> > be seen as not up to date with the competition and it will lose some >>> >> > of its popularity. And most people who use those libraries are not >>> >> > always that worried about the performance side of things, often it is >>> >> > more developer that tend to worry about that and frankly by my work >>> >> > experience I can say even in professional environment there isn't >>> >> > enough of them that actually care. You still have tonnes of companies >>> >> > doing web application that will not behave well outside IE, no >>> >> > validation, no multibrowser support, no javascript off support >>> >> > etc.... >>> >> > >>> >> > Personally I like to have my little benchmark and try to improve >>> >> > speed >>> >> > and all (mostly in php still, in javascript I still got a lot to >>> >> > learn, too often wrong :p) >>> >> > >>> >> > Hopefully things will get better in the future now that IE8 is out >>> >> > (also the number of current IE6 users still alarming), it's still not >>> >> > perfect of course, but the few things I tried on it so far behave >>> >> > more >>> >> > like they are suppose to. >>> >> > >>> >> > Wouldn't be nice when you can just code thing the way they should and >>> >> > they would work everywhere. Imagine the footprint of everything >>> >> > without all the lines for IE and other weird borsers lol >>> >> > >>> >> > >>> >> > >>> >> > On Apr 2, 12:32 pm, Andrea Giammarchi <[email protected]> >>> >> > wrote: >>> >> >> performances improvements and 44 seconds rather than 3 do not make >>> >> >> sense, >>> >> >> imho. >>> >> >> >>> >> >> What I am saying is that we, as libraries developers, just put >>> >> >> layers >>> >> >> to >>> >> >> make IE behaves like every other cool browser. >>> >> >> This layer at the same time is slowing down IE speed, which has >>> >> >> already >>> >> >> the >>> >> >> worse JS engine ever, compared to others. >>> >> >> >>> >> >> So, the equation is simple: cool stuff : IE = performances : bad >>> >> >> >>> >> >> I am sure a lot of code is focused just on the left part of the >>> >> >> equation >>> >> >> while should be nice, in my opinion, to consider more the right side >>> >> >> when it >>> >> >> is possible. >>> >> >> >>> >> >> Better? >>> >> >> >>> >> >> On Thu, Apr 2, 2009 at 12:20 PM, Gilles <[email protected]> wrote: >>> >> >> >>> >> >> > Déjá vus are actually well explained in the latest New Scientist, >>> >> >> > very >>> >> >> > cool stuff :) >>> >> >> >>> >> >> > I think you might want to highlight JQuery 1.3.2 and DOM instead >>> >> >> > of >>> >> >> > JQuery 1.2.6 tho as the performanace improved dramatically since >>> >> >> > then. >>> >> >> >>> >> >> > But I just don't get what you are saying by saying that libraries >>> >> >> > (not >>> >> >> > talking about JQuery only, cos I am only a user) are not optimized >>> >> >> > for >>> >> >> > IE. Any libraries will have a huge amount of lines that are >>> >> >> > dedicated >>> >> >> > to IE, cos IE just can't use standards properly. Unless of course >>> >> >> > as >>> >> >> > usual I don't get it cos the translation doesn't go well lol >>> >> >> >>> >> >> > If it wasn't for IE the number of check in libraries would be >>> >> >> > smaller >>> >> >> > and libraries would probably be faster lol >>> >> >> >>> >> >> > On Apr 2, 11:05 am, Andrea Giammarchi >>> >> >> > <[email protected]> >>> >> >> > wrote: >>> >> >> > > dejavou ... I am not blaming any library and obviously it is not >>> >> >> > > possible >>> >> >> > to >>> >> >> > > match manual tasks speed but Internet Explorer is not optimized >>> >> >> > > at >>> >> >> > > all, >>> >> >> > > that's the point. >>> >> >> >>> >> >> > > That TaskSpeed is performing truly common operation like remove, >>> >> >> > > append, >>> >> >> > > prepend, before, after, creation, etc ... and results against >>> >> >> > > other >>> >> >> > > libraries ( except for Prototype ) are clear in my blog post: >>> >> >> >>> >> >> > >>> >> >> > > >http://webreflection.blogspot.com/2009/04/taskspeed-dom-vs-libraries.... >>> >> >> >>> >> >> > > Here a summary of Internet Explorer results: >>> >> >> >>> >> >> > > Internet Explorer 6 >>> >> >> > > ------------------------ >>> >> >> > > library | ms >>> >> >> > > ------------------------ >>> >> >> > > DOM *3966* >>> >> >> > > Dojo 1.2.3 22750 >>> >> >> > > plugd-a (Dojo) 24189 >>> >> >> > > Dojo 1.3.0 24249 >>> >> >> > > jQuery 1.3.2 44314 >>> >> >> > > MooTools 1.2.1 69346 >>> >> >> > > jQuery 1.2.6 *104407* >>> >> >> >>> >> >> > > Internet Explorer 7 >>> >> >> > > ------------------------ >>> >> >> > > library | ms >>> >> >> > > ------------------------ >>> >> >> > > DOM *612* >>> >> >> > > Dojo 1.2.3 3641 >>> >> >> > > plugd-a (Dojo) 3748 >>> >> >> > > Dojo 1.3.0 3861 >>> >> >> > > jQuery 1.3.2 5031 >>> >> >> > > MooTools 1.2.1 7658 >>> >> >> > > jQuery 1.2.6 *10468* >>> >> >> >>> >> >> > > Internet Explorer 8 >>> >> >> > > ------------------------ >>> >> >> > > library | ms >>> >> >> > > ------------------------ >>> >> >> > > DOM *499* >>> >> >> > > Dojo 1.3.0 2327 >>> >> >> > > plugd-a (Dojo) 2455 >>> >> >> > > Dojo 1.2.3 2922 >>> >> >> > > jQuery 1.3.2 3170 >>> >> >> > > MooTools 1.2.1 5876 >>> >> >> > > jQuery 1.2.6 *7049* >>> >> >> >>> >> >> > > On Thu, Apr 2, 2009 at 10:49 AM, Gilles <[email protected]> >>> >> >> > > wrote: >>> >> >> >>> >> >> > > > It would be good to see DOM speed results compare to other >>> >> >> > > > libraries >>> >> >> > > > and see which libraries comes the closest to it. But I would >>> >> >> > > > never >>> >> >> > > > expect any library to match it on all tests. But then again >>> >> >> > > > that >>> >> >> > > > TaskSpeed test only take into consideration few basics tests >>> >> >> > > > and >>> >> >> > > > not >>> >> >> > > > the overall performance of akll the functionalities within a >>> >> >> > > > library. >>> >> > > >>> >> > >>> >> >>> >> >>> > >>> > >>> > > >>> > >>> >>> >> >> >> >> >> > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "jQuery Development" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/jquery-dev?hl=en -~----------~----~----~----~------~----~------~--~---
