Re: lynx-dev Java script support.

[Heather Stern]

Larry Virden >> I think, if I had enough need for Javascript that I was 
             >> going to add it, I would focus on some kind of neutral model 
             >> initially and worry about how to add such a model to the 
             >> currently object modeless lynx.  

David Wooley > That's not what people want; they want to use pages designed
             > for the big two.  

I have to agree in the case of Javascript;  in fact, not only are they more
interested in terms of the symptom (code for the big two works here) there
are a limited number of things that are common javascript uses, and if we 
recognized them only, many folk would still be pleased, I think.

I think these features are -
        * mouseovers 
          - some sites don't bother with ALT, but they do have
            a good mouseover status line
          - sites use mouseover on their nav to change a graphic to a
            a more "exciting" one when it's live.
        * other status line tweaks
        * a link which pops up a small javascript window instead of a whole
          new page
          - includes GeoCities popups, and many other ads. sigh
        * redirection
          - I think I saw one site use the javascript redirect acting
            differently from the plain text to "drop" the enduser into
            either the javascript'd area or the normal one.  But many
            many sites only have this type of redirect, not the Location:
        * CGI validation - the submit button invokes a javascript, the js
          does the "right submit thing" only if it's happy with the code.
          Without js on, these are broken forms.

Larry >> Isn't there either an ECMAScript object model or a W3C one that 
      >> one could use when adding scripting?

Rob Partington> Yes, the W3C DOM.  

(was to Larry)
David Wooley > Those pages are not written for the W3C 
             > object model, they are written for the IE4 or NS4, or if you
             > are lucky, the lowest common denominator of these two.  
             > Neither uses W3C DOM in its native mode.

Rob > That's what I've been working on for the last year or so, 
    > intermittently.  It's a requirement for any future versions of Lynx, 
    > IMHO, purely because it opens up so many avenues of future 
    > development -- not just Javascript, but Dynamic HTML, proper 
    > stylesheets, XML support, etc.

These are good things to want to support, but it feels rather like the
common problem when dining out in a large group - if the orders are too
complicated, they will be put in on all one ticket, and they won't bring
out *any* food until *all* the food is ready.  So the two out of the group
who ordered something small, get to starve longer than they had to, and
the waiter has to rush to get all the food out (having only two hands,
even if aided with a big tray)

Or said more succinctly, joe user just wants stuff to work, and isn't as
concerned with inner perfection, just outer noncrashing and "expected 
behavior".  (We could stand to improve the expectations so webmasters 
will write cleaner code, but that's a different direction.)

If we come up with some sort of "training model" whereby dev folk could work
on these great features seperately, then fit them to the truer model as you
get it completed, that would probably work.  At least the features would
gain by having some users trying to work with them.

Is the W3C model really so complex, or is it just hard to get time towards it?
If it is complex, no wonder the big two aren't using it natively.

My 2 cents plus inflation

-* Heather Stern -*- address@hidden -*- Starshine Technical Services *-