Re: lynx-dev Java Support

[David Woolley]

> Interesting.. I went to SureTrade to get some more specific examples, and
> here's a bit of interesting code...
> function IsNS3Browser() {
...
>         if (browserName == "Netscape" && browserVer >= 3)
>                 return 1 ;

You will find this sort of thing in any popular web page writing text
book (along with lots of other Lynx unfriendly stuff) and in the support
material produced by Microsoft and Netscape.  Basically most of the world
only cares about IE 4 and NS 4, and will do explicit tests for them,
rather than do feature tests in the Javascript or write to a lowest
common denominator.  (By feature tests I mean that it is often possible
to run some Javascript code in such a way that it will fail without an
error report if a required feature is missing, allowing a fallback
option to be invoked.)

> a lot.. and looking at the source, which is pretty C-ish, it looks like all
> it does is change which frame you're currently looking at based upon a bunch
> of criteria.. but there's only 3 frames on the 'parent' window..

It's becoming quite common to use Javascript to achieve something that
can be done with straight HTML.  E.g. I've seen several places where the
HTML is image maps pointing to variations of the same image map with a
different item highlighted, and which actually loads the required page
into another frame by means of a Javascript operation (my company's
marketing people have done this on the help pages for an application
that is otherwise fairly Lynx tolerant (but with no budget or mandate
for that tolerance)).

> 
> Anyhow, this kind of moving around documents is what I think is needed, 
> along with shopping cart kinds of uses (www.800.com).

I'm afraid my view is that the authors of such pages generally have no
interest except in supporting the latest IE and NS browsers++ and that
trying to track them with Lynx is a losing battle.  I would advise that
people should use NS or IE for such pages.

The people I have most sympathy for are blind users, but, at least in
the UK, no commercial web page designer is going to worry about them
until they become the last niche in a market that they have monopolised.
I'm somewhat sceptical that legislation will help here, unless it is
enforced; fire laws in the UK don't stop people jamming doors open with
fire extinguishers, because there is no realistic enforcement.

Incidentally, validation requires all the language and a large part of the
object model (except where the validation is bolier plate code from
an authoring tool, you cannot predict what language features and object
navigation paths will be used; in fact, validation is likely to use more
of the language than graphical gimmicks).

++ I can't even convince our marketing people and the other developers at
my employers to tolerate IE 3, let alone Lynx.  They only reluctantly
tolerate NS, as they think that NS will die.  This is for a product sold
for intranet use.  The problem is that software is sold on image, not
on function these days, so all the effort goes into getting the most 
visually attractive product for 80% of the market; it is very unlikely
that any of our customers use Lynx; there may be individuals, but they will
not affect purchasing policies.