W3C home > Mailing lists > Public > w3c-sgml-wg@w3.org > April 1997

Re: Error handling: yes, I did mean it

From: Sam Hunting <sgmlsh@CAM.ORG>
Date: Tue, 29 Apr 1997 04:25:44 -0400 (EDT)
To: lee@sq.com
cc: w3c-sgml-wg@w3.org
Message-ID: <Pine.GSO.3.94.970429041517.6116B-100000@Ocean.CAM.ORG>
> On error handling in XML...

For my money, the old CheckMark parser had both the best error
recovery strategy and the best user interface. An XML editor
that worked the same way would really win my heart....

It continuously validated the SGML instance, and when it encountered
an error it would (a) attempt to recover (mostly successfully), and
(b) put a vertical line on the status bar at the top of the screen
that showed roughly where the error was. So, one would end up with 
a status bar that looked like this, with white-space equalling valid
(heck, well formed) markup:

[   ||        |||||||||                      |     ]
    
    few...    big problem...                 just one...
   
So, one would click on the bar, that would scroll you to the code,
you'd get an error message, and you'd make the correction.

Often, in "big problem" areas, fixing the first error would cause
the rest of the errors to "uncascade", and one would experience
the extreme gratification of watching one's document seem to correct
itself, as the error bars disappeared one by one.

The only situation where the interface caused problems was when the
instance was so hosed that the bar was solid with errors -- and you
wouldn't want to be correcting something like that with an interactive
parser anyhow.

Sam Hunting
Received on Tuesday, 29 April 1997 04:25:51 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 10:04:25 EDT