Re: [xml-dev] Do you enjoy neighborhoods where every house looks the sam

["Jeremy H. Griffith" <jeremy@omsys.com>]

On Wed, 28 Aug 2013 15:31:09 +0000, "Costello, Roger L."
 <costello@mitre.org> wrote:

>Hi Jeremy,
>
>> we are basing it on MicroXML for simplicity ...
>
>This is fantastic. Is there a write-up where we can get 
>more information on this approach?

There will be; the draft spec is still a moving target.
I want to acknowledge that these recent threads have
provided strong validation for us, and are referenced 
themselves in the intro to the spec.

>> our first step was to write a fault-tolerant parser
>
>What is a fault-tolerant parser? What are its benefits?

The idea is that if the parser can tell what is missing
or incorrect, and correct it, it will do so, warn you, 
and proceed.  This is John Cowan's idea in the MicroXML
spec.  So for example if you have:

  <p>This is <b>bold.</p>

the parser can reasonably guess you meant:

  <p>This is <b>bold.</b></p>

Or if the character after the equals in an attribute
is not a quote, it can supply one.  There are quite a
few places the parser can make a reasonable guess.
Like, ",p>" might have been meant to be "<p>", with
the Shift key not in sync for the first character (a
personal favorite typo for me ;-).

In addition, in the next layer up, if we detect non-
whitespace text in an element that doesn't allow text,
we simply start a <p>, and end it at the end of the
enclosing element, or at the start of another non-text 
element.  Within a text element, a double return ends 
and restarts the current element.  This policy greatly 
decreases the tagging a writer needs to do.  The writer 
can just write, which is, after all, her real job.

>Is it publicly available?

It will be.  We don't want to start it off with any
features we will want to change the next day, and 
that is still what we are doing now.  But soon.  <g>
 

-- Jeremy H. Griffith <jeremy@omsys.com>
   DITA2Go site:  http://www.dita2go.com/