Re: LYNX-DEV Lynx v2.3.6 and poll submit problem

[Klaus Weide]

[I am cc'ing this message to lynx-dev since somebody else may want to
 comment]

On Tue, 15 Oct 1996, Alicia Perdue wrote:

> Date: Tue, 15 Oct 1996 22:58:15 -0400 (EDT)
> From: Alicia Perdue <address@hidden>
> To: Klaus Weide <address@hidden>
> Subject: Re: LYNX-DEV Lynx v2.3.6 and poll submit problem
> 
> 
> On Sun, 29 Sep 1996, Klaus Weide wrote:
> 
> > On Sat, 28 Sep 1996, Alicia Perdue wrote:
> > > I am having a problem with Lynx v2.3.6 and submitting my vote in polls
> > > such as the one featured @
> > > http://www.nascar.com/fans/polls/0923tysonpick.html
> 
> >  I think you have got the version numbers wrong.  You must mean 2.5 and
> > 2.6, without the middle number.
> 
> Yes, I meant v2.6.
> 
> >  Anyway, yes the behavior of Lynx has changed between versions 2.5 and
> > 2.6.  Lynx is now doing "the right thing" (following the specs of the
> > HTTP protocol), while many other browsers don't.
> 
> <interesting tech explanation snipped> =)
> 
> >  So Lynx (now) is doing what the protocol requires.  It assumes that
> > the page author knows what he/she is doing.  The page asks for a "POST"
> > submission, the servers then asks for redirection of the "POST" request,
> > and that is what Lynx will do.  It will, however, first ask the user
> > for confirmation, before re-issuing the request:
> 
> >  So why are page authors doing this?  The most likely answer is that they
> > have never encountered a browser which does the right thing.  They have
> > probably only experienced browsers which silently converts a "POST"
> > into a "GET" upon redirection.
> 
> I guess my question is why did you guys change it?  I know and respect the
> fact that the lynx dev. team follows HTML standards faithfully.  But, as
> an end-user, I feel caught up in the battle between M$/NS non-standard
> HTML and lynx.  What once worked for the end user now doesn't.  I really
> can't switch back to v2.5 since I like the frames-support of v2.6.  Is
> there any purpose (other than making a statement on staying faithful to
> the HTML standard) for changing how Lynx handles this?
> 
> Thanks,
> Alicia <address@hidden>
> 
> ps: sorry for the long delay in the reply.  I had saved it but never
> remembered to reply!

Well, you didn't dispute the statement that it was "the right thing"...
So I assume you accept that, and aren't those three words justification
enough? :->

But anyway, I think at the time the change was made in Lynx nobody of
the developers knew just _how_ common this (mis)use of redirection was.
There is a standard, and so Lynx was changed to behave according to the
standard - it seems like a reasonable expectation that that would be
a good thing.  After all, all those other software products are claiming
to follow the same standards...  

I don't know whether the "old" Lynx behavior was the result of a 
conscious decision, or whether it just was the way it was because
nobody had thought about it a long time ago.  The part of the code that
(mostly) deals with this belongs to the CERN (now W3C) Reference Library
portion, on which (I believe) also early versions of Mosaic and Netscape
were based, so it's no surprise that these browsers behaved in a common
way.

Anyway, since it has became obvious that this change left a number
of users "caught up in the battle" (as you say), somebody has made
a patch that will allow users to choose the old behavior.  (There still
will be a message and a prompt, which you may find annoying. But IMHO 
it is a good thing to make users aware that they are dealing
with a web site that does not understand the protocols :-) )
That patch has not made it into any new release of Lynx (there haven't
been any), but it probably will.

You asked whether there was a purpose (other than making a statement)
to the change.  Well first, this is about being faithful to the HTTP,
not HTML, standard - which is more basic, and IMO more important to
adhere to.  Among other things, there *is* something that resembles
a real standard, which cannot so easily be said about HTML.
And there *are* good reasons for what the standards say - the folks
who wrote them have thought a lot about the issues, and their intention
is definitely not to make life hard for users or to spite Netscape etc.

  Klaus

;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;