Comments on: JavaScript optimization, are chained calls expensive? http://dossy.org/2006/03/javascript-optimization-are-chained-calls-expensive/ Everything that comes out of Dossy, from the strange to the banal. Fri, 06 Nov 2009 22:44:13 -0500 http://wordpress.org/?v=2.8.5 hourly 1 By: sree http://dossy.org/2006/03/javascript-optimization-are-chained-calls-expensive/comment-page-1/#comment-346 sree Fri, 10 Mar 2006 03:32:49 +0000 http://dossy.org/archives/000248.html#comment-346 Hey Dossy, I hadn't finished part 3, where I get into the numbers, but I'd think 30% is still a pretty big deal, IF you're writing applications at scale - in the trivial case, totally agree with you - who cares - that's indeed part of the point of Javascript. Also, dom/activex/livescript nested references are even slower than "native" javascript object ones - so that can have more of an impact. Additionally, none of the browser javascript interpreters can do squat meaningfully to optimize (even JScript.NET, as I'll show in pt 3) because unfortunately, Javascript is inherently a dynamic language - the function invocation (or even property invocation) may cause volatility (e.g. self-modification) with respect to the calling function/object. Doesn't mean you CAN'T deal with it, but none do. Incidentally, though, I wasn't trying to make a big deal out that optimization, per se - I was only trying to suggest that understanding the WHY of the perf delta will help you understand the value of other language features like the "prototype" property. Hope that helps clarify :) Hey Dossy, I hadn’t finished part 3, where I get into the numbers, but I’d think 30% is still a pretty big deal, IF you’re writing applications at scale – in the trivial case, totally agree with you – who cares – that’s indeed part of the point of Javascript.

Also, dom/activex/livescript nested references are even slower than “native” javascript object ones – so that can have more of an impact.

Additionally, none of the browser javascript interpreters can do squat meaningfully to optimize (even JScript.NET, as I’ll show in pt 3) because unfortunately, Javascript is inherently a dynamic language – the function invocation (or even property invocation) may cause volatility (e.g. self-modification) with respect to the calling function/object. Doesn’t mean you CAN’T deal with it, but none do.

Incidentally, though, I wasn’t trying to make a big deal out that optimization, per se – I was only trying to suggest that understanding the WHY of the perf delta will help you understand the value of other language features like the “prototype” property.

Hope that helps clarify :)

]]>