[XSL-LIST Mailing List Archive Home] [By Thread] [By Date]

RE: Style vs. transformation


Subject: RE: Style vs. transformation
From: "Reynolds, Gregg" <greynolds@xxxxxxxxxxxxxx>
Date: Wed, 4 Mar 1998 16:42:39 -0600

> -----Original Message-----
> From:	Darrin Smart [SMTP:darrin@xxxxxxxxxxxxxxxxxxxxx]
> Are there any free implementations of JavaScript (with source)?
> 
Netscape Navigator?  Well, real soon now.

A naive question (I'm shaky on interpreter implementation):  Why not
specify the scripting language abstractly, as a collection of functions
and datatypes.  So instead of stipulating, for example, that the
addition operator is infixed '+', you stipulate that the addition
operator (whatever it looks like) applied to numeric args sums them and
returns a number.  Of course you need a reference concrete language (is
this too SGMLish?), but conformance would be defined in terms of the
abstract grammar and semantics.  In other words, why not specify a
meta-language from which implementation languages "inherit".  

I know the obvious objection is that "we" want a one scripting language
that will work everywhere.  But  how hard would it be to then support
multiple languages? (The foregoing is not a rhetorical question.)
Wouldn't it just be a case of parsing the source, tranlating it into a
native form with simple table lookups, and proceding?  Speaking as an
unreasonable user (writer of scripts) I demand Scheme.  But being an
equally unreasonable browser, I demand support for JavaScript, Python,
and whatever other language web authors may choose.  So let the market
decide - it's virtually certain that all vendors would stick with
javascript, but who knows?  I would spend my money on a more versatile
browser.


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list



Current Thread