Re: lynx-dev JavaScript again (was no subject)

[Klaus Weide]

On Fri, 22 Oct 1999 address@hidden wrote:
> In message <address@hidden>, 
>            Klaus Weide writes:
> > On Wed, 20 Oct 1999 address@hidden wrote:
> > 
> > > Lynx has no internal document structure
> > 
> > That alone seems to make javascript (as usually used) pretty much
> > pointless for lynx as it is.
> 
> On the whole, yes, it does.  :-(  However, it's still possible to use
> Javascript'd links and "correct" form validation routines.  
> 
> As regards Lynx's lack of internal document structure, it's not difficult
> to build something that creates an object tree *if* you have well-formed
> documents or an error-correcting parser.  If SortaSGML can be extended to
> become error-correcting, then I have the object tree code already.  Then,
> all that needs doing is to figure out how to render the object tree.

What errors need to be corrected?  What kind of guarantees do you need
about the "SGML" parser's output?

> I did have the idea of replacing Lynx's internal "objects" with Javascript
> classes and objects.  It simplifies things, not only with regards to 
> Javascript, but with regards maintainability et al.

So how would lynx access its objects after you have "taken them away"?
They're C, now, if you make them "Javascript classes and objects" they
aren't...

  Klaus