Re: LYNX-DEV Patch for development version 0.58

[Klaus Weide]

On Wed, 3 Sep 1997, Doug Kaufman wrote:

> On Wed, 3 Sep 1997, T.E.Dickey wrote:
> 
> > you wouldn't be able to search for a "word" except by typing "drow", which
> > some people might object to.  (And since, as I understand it, rendering
> > is in terms of strings, I'd suppose the whole page would be mirror-image
> > when displayed).
> 
> As I understand it, the character sets specify which pattern will be
> displayed for each character _individually_.  I don't see where
> left-to-right or right-to-left is involved with individual character
> representation on the screen.  Searching for a latin word should work
> fine.  Searching for a word written right-to-left would be more
> difficult, and would probably be best done by printing the web page and
> using a tool designed for right-to-left languages, but as you
> demonstrated, it isn't that difficult to write "drow".
>                               Doug

The text is transmitted and stored in logical order, not in visual order -
at least sometimes, and according to Unicode standard and to RFC 2070
(see 4.2.4 there) and to the HTML 4.0 draft.  I can't really tell whether
that is the case for the pages I looked at, because I cannot read a word
of it (so I can't tell whether I am reading a "word" or a "drow"), but at
least for newer Web pages I think that is always the case.  For mail it
may be different.  There is actually defined iso-8859-8 and iso-8859-8-i
and iso-8859-e, depending on whether "visual" or "implicit" or "explicit"
directionality is used  (RFC 1556.  And it gets really fun when
left-to-right and right-to-left text parts are mixed.

So basically Lynx will show text written in Hebrew and Arabic reversed.
I don't know whether users of those languages are more trained than others
to read sdrawkcab, maybe they are more used to this kind of trouble.
WIthout *lots* of changes lynx support for these scripts will be rather
half-assed, but since it seems Fote and you would like to see it - well
the more the merrier. :)

Newer pages (including those created by Alis's Tango) seem to prefer
Mircrosoft or maybe Macintosh encodings for Hebrew and Yiddish (according
to my limited looking around), which are more complete than the minimal
iso-8859-8 character repertoire, so then we should probably add those
also.

    Klaus


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