Comments on: Microsoft IE JavaScript Perf. Tips, Part Trois

By: Klaus

Klaus — Tue, 09 Jan 2007 10:22:43 +0000

A properly implemented garbage collector can collect closures without any problems whatsoever.

By: Donald Duck

Donald Duck — Mon, 08 Jan 2007 18:55:29 +0000

IE code fix:
if(isIE){
alert(“please download firefox”)
return true;
}
//problem solved!!!

By: Valery Silaev

Valery Silaev — Mon, 08 Jan 2007 10:34:47 +0000

@Alex Russell
—88—-

Can’t stress this more. Just imaging same sentece from authors of Scheme or Haskell interpreter: “Tail recursion are both extremely powerful and extremly dangerous” :))

By: Valery Silaev

Valery Silaev — Mon, 08 Jan 2007 10:28:29 +0000

Unless MS fix their terribly slow DOM all their “performance tips” are irrelevant. Any optimization starts with analysis, afterward the slowest parts of system are fixed. In case with MSIE the only optimization possible is to not use MSIE or stick with innerHTML for another 5-7 years till MS will understand the need to buy someone’s else HTML rendering engine.

By: Alex Russell

Alex Russell — Sun, 07 Jan 2007 05:44:52 +0000

I don’t know what’s more embarrassing for the JScript team: they their interpreter and its assumptions are this broken or that they’re not ending each of these articles with an apology to the world.

If they were to be really honest with us, each of these would begin with “we’re sorry we screwed up this badly, but it’s deployed and that’s just how it is for now. Here’s what you can do about it…”.

That the article includes the sentence “Closures (…) are both extremely powerful and extremely dangerous” should be the tip off to everyone that the problem isn’t with our code, it’s with their engine. They may as well be telling people not to write OO code at all.

C’mon MS. You can, and should, do better.

Regards

By: Adam Sanderson

Adam Sanderson — Fri, 05 Jan 2007 17:53:06 +0000

Wow, if I don’t use closures I might as well not right code at all. It’s good to know where the inefficiencies are though.

By: Dean Edwards

Dean Edwards — Fri, 05 Jan 2007 15:38:11 +0000

What’s funny is that the first and second tips are the same. The property accessors defined in the constructor are slower because they are closures.

By: Mark Kahn

Mark Kahn — Fri, 05 Jan 2007 15:29:20 +0000

from the article:

“perfCar.m_tireSize = iTireSize + 1; // An upgrade”
// [is preferable over]
function Car_put_tireSize(value){
this.m_tireSize = value;
}
ooCar.SetTireSize(iTireSize + 1); // An upgrade

… and in the example of how things “should” be done:

PrototypeCar.prototype.GetTireSize = function() { return this.m_tireSize; };
PrototypeCar.prototype.SetTireSize = function(value) { this.m_tireSize = value; };
// …
protoCar.SetTireSize(protoCar.GetTireSize() + 1)

I love it!

By: Jack Slocum

Jack Slocum — Fri, 05 Jan 2007 14:31:19 +0000

Judging from the quality of the article code samples (1, 2, and 3), the IE Team choose the wrong guy for this series of articles.

By: Giovambattista Fazioli

Giovambattista Fazioli — Fri, 05 Jan 2007 13:35:08 +0000

Microsoft has a hilarious politic. They are proud for JScript, but they recommended some of very strange. Why I can not use object oriented programming like get and set? Are they crazy! So, come back to QuickBasic…

By: Morgan Roderick

Morgan Roderick — Fri, 05 Jan 2007 09:05:19 +0000

The MS article itself is of questionable value, urging programmers to create “funky” scripting to satisfy IE’s poor implementation of javascript / dom / whatever. General articles on creating high quality scripting would be of much greater value to the community, and might even get MS some respect from the community.

The comments on the article on the other hand are hilarious!