Re: lynx-dev Lynx 2.7.1 and 2.8 refuse to render certain HTML documents

[David Woolley]

> Also, the current design has the dilemma that if you need to re-render a
> reply to a form submission, you can't do it without resubmitting the form=
> ,
> and resubmission may not be safe, so users must be prompted for a decisio=
> n
> on whether to do it, based on considerations that they generally don't
> understand.  Nothing about Bela's suggestions precludes not cashing

That's a good point, but would it have been made explicit if I hadn't been
challenging the ideas?

> insisting Lynx only be able to do what is optimal in you own usual
                                                               ^--- un ???
> circumstances?  (There are a lot of things about Lynx that have become
> unattractive during the past year and a half, and more are likely to
> follow as things presently stand, but your assessment of the circumstance=
> s
> under which Lynx "tends" to be attractive is very limited, and what you'r=
> e
> insisting on amounts to a self-fulfilling prophecy for it.) =


I think there is a trend to try to make Lynx compete with commercial GUI
browsers, which I don't think will succeed until, like black and white 
material in advertisements, character mode suddenly becomes fashionable again
because GUI has become boring.  As such it should try to play on its
strengths, which include low resource requirements.  Although I didn't like
the style of the article, I think there is some truth in the claim that
Lynx is elitist in the rich parts of the world, and I think that the text
mode behaviour is key to that.  I have to admit, that in the office, where
I have the processing power to run MSIE, I use it except for diagnostic use
and problem pages as so few of the sites that I need are optimised for 
text mode and several won't even work with MSIE3.  (In private life, one
can chose to ignore vendors with unuseable pages, but as an employee,
one can't make such policy.)

(Wearing a producer hat, I also know that businesses don't care about
(read: "it is commercially naive to worry about") useability of pages on
anything except the browser that account for 80% of their target market.
I know of HTML being created by people who barely understand it, but
they are creating it well enough to meet the marketing requirements, so
one can't force them to do things properly, e.g. CSS instead of deprecated
attributes.  There are some businesses for which this is not true,
because they have already saturated the MSIE/Navigator 4 market and
are big enough to have a return on investment for making the pages more
generally useable, but that is only a few.  The consequences of this are
that pages that are essentially GUI only will continue to be produced,
even after HTML is extended to allow the desired effects to be achieved in
compatible ways, and therefore that Lynx is always going to be difficult
to use by mass market users on mass market pages.)

The other problem with the continued bloat is that fewer and fewer people
have the time and ability to understand Lynx well enough to maintain it.
That's not so much of a problem for someone like Microsoft, who can have
a formal team with one person who understands the overall structure and
others who are expert in one piece, but decentralised maintenance tends
to limit updates to those who have a reasonable understanding of the
whole of the program.

Incidentally, of your other concerns, I think internationalisation is a
valid area for Lynx, although I expect few sites to actually correctly
support it, but, I think that tag soup will only be a short term problem 
unless CSS and dynamic HTML fail in the marketplace.  CSS, I think Lynx 
should do; dynamic HTML is likely to be problematic, because I think 
designers will use it in ways which cannot be represented sensibly in text
mode.  On SSL, I don't object to the concept, I object to what I consider
the legally unsafe positions being taken and the consequent risk that 
breaches will be used as the basis as of a fear, uncertainty and doubt
attack on free software, and on the credibility of the Gnu licence.

In the end though, you all have to make your own minds up.  Ideally,
writers should state facts and their own opinions, but be very careful
to distinguish between errors or incompleteness in facts by others, on
one side, and differing value judgements and opinions on the other side.
Finally, what actually gets done will be the decision of those motivated
enough to spend significant time on Lynx, and may well represent their own
pet concerns or learning wishes.  I openly admit that I don't have the
time to do anything significant on Lynx, other than to try to contribute
to the debate, although I believe that can be valuable, even if things
don't go my way, as long as people think about what I say.