Can't help but find this a bit discouraging. Modules, yes, but it does seem like they do want to inject everything in there (including the kitchen sink), without actually fixing some of the more fundamental language flaws (little things like true/false/null being constant keywords, and undefined being a variable).

I mean, I don't even hate the arguments object that much... It would actually work O.K. if the spec didn't have the insane requirement that the parameters must alias with the object fields, and that the object isn't even a real array...

I'll try to relax and remind myself that their spec is due for 2013, which probably means 2014-2015, and that I might be getting pretty close to done with my Ph.D. by then!

You know, Marc, I think your class is giving me more appreciation for Scheme. It feels pretty robust, as a language. Now if someone could just find a way to solve the parenthesis problem ;)

- Maxime

Then maybe I can do research on Scheme.

On 10-11-25 10:48 AM, Bruno Dufour wrote:

On 2010-11-25, at 24:15 , Marc Feeley wrote:

...

Brendan Eich gave a presentation on the next spec for JS:

http://hacks.mozilla.org/2010/07/brendan-eich-at-jsconf-2010-whats-coming-in...

It fixes many of JS's problems. It also looks more and more like Scheme! (no parens... but rest parameters, let, continuations, weak references, modules, ...)

So, it seems like we don't have to worry too much about optimizing when programs use the arguments array. But we might have to worry about JS getting the kitchen sink too in a near future :) Also, the last few minutes of the questions are really interesting (the part where he talks about how JITs behave when their assumptions break...)

Bruno _______________________________________________ Tachyon-list mailing list Tachyon-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/tachyon-list

...

I'm not sure it is a good idea to strive for instant irrelevance

even though that seems to happen often with PhD's...

If the new standard comes out slightly before or after I'm done with my thesis, I don't think that will make my work irrelevant. I mean, I thought I was doing a thesis on optimizing dynamic languages (which ES5 happens to be one), not specifically on JavaScript/Web 3.0.

If we can find time to slip in features like const support, modules, etc. I won't oppose to it, so long as it doesn't end up taking most of our time and delaying research work.

...

There are two ways to solve that problem...

I was actually thinking there might be a third way. A better visual editor that recognizes special forms, offers color coding, and possibly does indentation for you. Something that makes the parentheses outright invisible.

- Maxime