Re: LYNX-DEV why not use shift-V to access multiple bookmarks list?

[Benjamin C. W. Sittler]

On Thu, 7 Nov 1996, Al Gilman wrote:

>   From: address@hidden (Larry W. Virden, x2487)
>   
>   What I mean is that in some manner, either at compile or run time, we
>   should consider providing a way to make lynx as easy to use for the
>   speech synthesisers as possible - if we are going to use this segment
>   of the internet as a bargaining chip when harassing sites to make them
>   lynx compatible.
> 
> There are two strategies I have been mulling:
> 
>       - define a capability-envelope for a virtual terminal class
>       that at least 80% of the screen readers can use; use the
>       TERMCAP/TERMINFO mechanism to tell lynx that this is what you
>       have; let curses handle it.  This is eminently doable.  If
>       for example Gregory and Patrick J. Burke agreed to lead a caucus in 
>       lynx-learners to define the virtual terminal, that would get
>       done, and then they could explain what to do with the
>       capabilities of that terminal.  This already would be a         
>       significant positive step.  

Perhaps instead of defining this through a TERMCAP/TERMINFO entry we could
bypass curses altogether for speech users. Wouldn't an interface closer to
the old CERN linemode browser be preferable? Of course, if we did stick
with curses we could abort rendering immediately when a link is selected, 
since we'd be in cbreak mode. Actually, we could use cbreak() without
bothering with the rest of curses, and have our output look something like
this:
Title: document title
Heading: document title
Subheading: first section
Para: Entire text of first paragraph, with appropriate character sequences
for emphasized text and with all links [1] numbered in a style usable by
the synthesizer. Rendering would be aborted as soon as a link was
selected.

>       - [more work] Fote talked about six styles used in rendering
>       UL.  If you study the problem, you realize that you want to
>       tear up the display algorithms clear back before curses and use
>       a different library of styles to present an HText in speech
>       than on VT-100.  For example, a SELECT feature in the HText
>       would be managed in the user dialog much the same as a 
>       vertically-stacked column of RADIO button options, where the
>       speech user is able to index through the options and hear one
>       at a time (possibly with pop out to wrapper, pop back to 
>       current index).

How about rendering selection lists like this:
Please select your browser from the following:
Lynx [1]
Emacs W3-mode [2]
Agora [3]
DOSLynx [4]

and allowing rendering abortion whenever a link is suggested. Yes, this
would require that Lynx output in small chunks and keep checking for
input, but it might be worthwhile. 

> That's a dump of currently resident confusion on speech-friendly webbing
> in this noggin.

> Al Gilman

Same here... any speech users want to comment on this?

--
Benjamin C. W. Sittler - address@hidden - http://www.nmt.edu/~bsittler
#!/bin/perl -sp0777i<X+d*lMLa^*lN%0]dsXx++lMlN/dsM0<j]dsj
$/=unpack('H*',$_);$_=`echo 16dio\U$k"SK$/SM$n\EsN0p[lN*1
lK[d2%Sa2/d0$^Ixp"|dc`;s/\W//g;$_=pack('H*',/((..)*)$/)

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