Re: lynx-dev SWITCH_DTD help text

[Klaus Weide]

On Sun, 27 Dec 1998 address@hidden wrote:
> In a recent note, Klaus Weide said:
>
> > I don't like calling this "HTML error tolerance".  [...]
> > 
> The name works if you understand it to indicate not degrees of
> tolerance, but alternative modes of tolerance.

Ok - I still think "error tolerance" is just a side effect, but anyway..

How about this text?

------------- snip -----------------
Lynx often has to deal with invalid HTML markup.  It always tries to
recover from errors, but there is no universally correct way for doing
this.  As a result, there are two parsing modes:
"<DFN>SortaSGML</DFN>" attempts to enforce valid nesting of most tags
at an earlier stage of processing, while "<DFN>TagSoup</DFN>" relies
more on the HTML rendering stage to mimick the behavior of some other
browsers.
You can also switch between these modes with the CTRL-V key, and the
default can be changed in lynx.cfg or with the -tagsoup command line
switch.

<P>
The "SortaSGML" mode will often appear to be more strict, and makes
some errors apparent that are otherwise unnoticeable.  One particular
difference is the handling of block elements or
&lt;li&gt;..&lt;/li&gt; inside &lt;a HREF="some.url"&gt;..&lt;/a&gt;.
Invalid nesting like this may turn anchors into hidden links which
cannot be easily followed, this is avoided in "TagSoup" mode.  See the
<a href="follow_help.html">help on following links by
number</a> for more information on hidden links.  Often pages may be
more readable in "TagSoup" mode, but sometimes the opposite is true.
Most documents with valid HTML, and documents with only minor errors,
should be rendered the same way in both modes.

<P>
If you are curious about what goes on behind the scenes, but find that
the information from the -trace switch is just too much, Lynx can be
started with the -preparsed switch; going into SOURCE mode ('\' key)
and toggling the parsing mode (with CTRL-V) should then show some of
the differences.