On 12 January 2014 08:54, Brendan Eich <[email protected]> wrote: > Not on the (gzipped) wire: > http://mozakai.blogspot.com/2011/11/code-size-when-compiling-to-javascript.html
I am not worried about the compressed size but the latency caused by decompressing and parsing the contents. Actually I am not worried at all - this is not my complaint but one I've heard from others. I mentioned it as one example. I don't plan to use Asm.js, but potential users might be unsettled by how unoptimal the lexical representation is, even if it is only "theoretically unoptimal". I've heard that there are plans to include a special case parser Asm.js in SpiderMonkey, is this true? I can see how for many onlookers this is the point where they will ask: why don't you specify a bytecode? Which brings me to what I said in the original post: introduce the capability to convert non-JS code fast, and onlookers' misgivings about using JS syntax as a transport might go away. And we don't even have to design a bytecode up front. _______________________________________________ es-discuss mailing list [email protected] https://mail.mozilla.org/listinfo/es-discuss
