Re: lynx-dev TRST: the next step

[Klaus Weide]

On Sun, 7 Nov 1999, Philip Webb wrote:
> 991106 Klaus Weide replied to Philip Webb:
> > An excellent job?  Not exactly.
> > Did you notice the difference when toggling ^V?
> 
> no: why should i toggle ^v ?  

Uh, to notice the difference I am talking about, maybe?
To verify that the HTML is broken enough that it makes a difference?

> it does look worse with 2-8-2dev.14 .

What looks worse than what??  You are not making sense to me.

> the secretarial staff at the TTC have no expertise in HTML,
> little time to spare for the conversion exercise, no money to waste
> & no time to spend explaining to me exactly what they do;
> they have probably checked the result with Notstraight & Exploiter
> & would be quite reasonable in asking me why i don't use one of them.

I hope you'd have a good answer. :)

> > If you want to work on this,
> > it might be useful to start with less broken examples.
> 
> this is not an idle "example": it's a very serious document,
> outlining capital spending on our transit system 2000-2009,
> whose content i want to read for professional purposes
> & its tables are very typical of those in such documents.
> Lynx should be able to reproduce those tables adequately.

We haven't found the perfect cure for GIGO yet...

>From a technical standpoint it *is* G, no matter how serious the
contents.  It's just not obviously G if viewed with a specific garbage
processor, because it was tailored for that (class of) G.P.  Of course
there are ways to more-or-less deal with that, which is what lynx
does.  I was just suggesting that you might want to use less peculiar
specimens for testing your modifications (and understanding the
results).  You know, start with the simple case and work up to more
complicated ones, rather than the other way.

> i notice you don't deny the original Lynx rendering is grotesque (wry grin).

I don't deny you your right to colourful expressions. :)

> what i want is TableSoup, preferably available by keystroke like Tagsoup.
>               
> > He even thinks he has some valid excuses for not starting down that road...
> 
> there are always excuses, but yours sound to me like the pedantic fatalism
> which i hoped had disappeared from Lynx development when FM retired (smile).

To be replaced by... Soup-of-the-day hackery? :)

FM kept lynx alive, cleaned up the whole mess, made numerous pragmatic
compromises.  Call that pedantic fatalism if you like.  I think you need
a better dictionary.

Well I guess I just don't have a "very serious document" that I would want
to adapt lynx to whatever it takes.  Call that pedantic fatalism if you
like.

> <BR> & <DIV> should also be ignored in TableSoup mode;
> i'm not sure about <CENTER>, which would depend on experience.
> 
> > What if the contains more than one <P>?
>              ^?
> ALL <P>'s would be ignored in TableSoup mode.

As I said, I am curious where you take this.  I hope you will do
some testing on a variety of tables.

> you did start to get tables working in Lynx:

(Depends entirely on your definition of "works".  Minimal table support
worked, too.)

> don't be so fatalistic about what can be achieved.

Did I forget to mention for a while that TRST is Not Full Table Support?
And that it is extremely unlikely that it can be the basis for that?

Sure, all kinds of things can be achieved; this may just not be the right
way.  But of course I don't have to be the achiever - glad to see you join
in...

   Klaus