Re: lynx-dev patch

[Vlad Harchev]

On Mon, 12 Jul 1999, David Woolley wrote:

> > 
> > 2) Seems lynx is now turned into good html->txt translator (it now justifies
> 
> This is a subjective statement by the author written as though it were
> an independent review.

> >    text and emit backspaces)! The justified ouput looks very nice for
> 
> That makes it a bad html to txt convertor, as the text is now difficult
> to post-process - it is essentially a final form document.  For printing
> purposes, the existing html2ps utility would be a more appropriate choice,
> for producing final form documents.

 Emitting backspaces is turned off buy default (it's turned of if
-with-backspaces switch is given), so no backward compatibility is
broken and no flexibility is lost. 
 Lynx documentation project uses plain text files with backspaces very
actively - this is one of the format Linux-HOWTOs are shipped in with almost
each linux distribution, but alas they are non-justified. So lynx can be used 
exactly for this purpose by LDP. And seems that no other html2txt converter
does (tableless) html->txt conversion so good except lynx, that's why note
number 2).

> 
> >    documents that contain human words (ie not unspittable URLs). 
> > Justification
> >    won't happen for CJK texts  ( -justify -dont-wrap-pre -with-backspaces).
> 
> Actually, final form documents are when you do want to wrap pre, or have
> it fall off the edge of the paper.

  Not wrapping in pre gives you ability to truncate <pre> texts rather than
fold them.

> > 6) From now seems support for lynx style sheets is very desirable (ie 
> > ability
> 
> What is needed is support for CSS style sheets (which incidentally
> control jusitifiation).

 I wrote in the previous message my POV on it - implementing exact semantics
of W3C can be not needed.

> >    to control indents of various text styles) - since justified text looks
> >    MUCH wider than non-justified. IMO the width of 'man' output is good.
> 
> This is one of the reasons why justification is not normally reccommended
> on the sort of short lines that you get on a text mode display.
 
> > begin 644 patch.gz
> > M'XL("`QGC#<``W!A=&-H`.P\:W/:R+*?X5?,2:IBS"N(AY]QSF*0`[L8?'C$
> > M2>VF5+(8C&(AJ201V^><W-]^NWM&TN``L9US;^W>&V^M(VFF>[I[^JV1I_9L
> 
> I'll have to see if I can find time to decrypt this, but things to
> check for are that justify (= style="text-align: justify")
> interacts correctly with CENTER and <DIV align="center">
> (= style="text-align: center").  Incidentally, many web pages are
> centered, which means that they cannot also be justified.
> (This centring is also because the author likes it.)

 This patch won't try to justify centered or rightaligned texts. If you noted
the way the centered pages are centered - what it is (wrapping entire file in
CENTER, using <p align=center> or <div align=center> or stylesheet? Seems that
for <p align=center> the lynx can be hacked in order to do justificaion even
inside those paragraphs. 

 Best regards,
  -Vlad