Re: Hello World problem
On 12/4/2012 6:09 PM, Arne Vajh??j wrote:
On 12/4/2012 5:43 PM, Martin Gregorie wrote:
The ECMA-262 specification was never mentioned on that newsgroup or I'd
have grabbed and used it.
It is pretty compact.
188 pages to cover both language and some core classes (a la
java.lang package in Java),
That's a pretty unfair comparison. JavaScript is dynamically-typed,
which means it just has to cover runtime semantics and can ignore a lot
of the complexity that statically-typed object-oriented systems bring.
Effectively, JS only needs to specify equivalents to sections 1-3, 14,
and 15 of the JLS (with small portions of other sections), which comes
out 215 pages by my count. Even then, the JLS also includes bulky parts
in these sections that JS can happily ignore a lot of, including how to
find candidate method overloads (a thick 30 pages unneeded in dynamic
languages).
The core JS library is also pitiful by Java's standards, being far
smaller than even just java.lang: Array corresponds vaguely to
java.lang.ArrayList, String to java.lang.String, Math to java.lang.Math,
Number to java.lang.Double, Date to java.util.Date (minus almost all
locale concerns), RegExp to java.text.regex.Pattern, and Object and
Function both covering features which don't really exist in Java but
could be considered roughly equivalent to java.lang.Class.
Runtime environment is a bit more tricky, because it somewhat depends on
the runtime.
Yes, sure. But, is that poor language design and specification or merely
a lack of rigour on the part of the interpreter writers?
Well - I am not a JavaScript expert, but based on my very limited
knowledge about this then the main problem is that use of JavaScript
in browsers are highly dependent on the HTML DOM model and CSS support
used in the browser. So JavaScript is really suffering from the chaos
in the HTML world and not so much due to its own problems.
When most people talk about JS, they're really talking about JS +
"HTML5", which I'm using here to refer to both the DOM APIs presented by
the HTML specification itself and the large collection of auxiliary APIs
that are assumed to be implemented by web browsers. Most variance in
practice tends to be associated with the DOM.
That said, there are some places where I've heard that increasing
compatibility is contentious by various engine writers; the biggest one
I know of is the iteration order of object properties (particularly if
you have both "foo" and 0 as properties...).
--
Beware of bugs in the above code; I have only proved it correct, not
tried it. -- Donald E. Knuth