Lists Home |
Date Index |
A DTD can be written to do exactly that by requiring an
acknowledgement of the warning. This is IETM markup
and that can be done.
But the point here is that different systems require
different amounts of validation and error correction
at different stages of processing. You have taken
a pretty hard line stance against one of the more
practical applications of context-aware editing.
The one worthwhile point is that this should
be able to be turned off so an author can
create ill-formed markup at points, or it
should be an on-demand operation as it is
in IADS and has been for a long time.
BTW: I attended the IADS conference briefly
last week. The thing is still pulling trains.
Amazing. It is worth a look for those who
want to play with a pre-web SGML-capable
browser and authoring tool and see what is
possible when one strips down to basics.
They offer XML support and URLs now. Wow.
From: Elliotte Rusty Harold [mailto:firstname.lastname@example.org]
At 3:58 PM -0500 4/9/04, Bullard, Claude L (Len) wrote:
>Yep. If a system lets a repair technician skip
>a warning or a writer inserts a caution instead
>of a warning, the repair technician can die.
Which has nothing to do with what we're discussing here. No software
system I've seen is smart enough to tell whether a particular problem
should be a caution or a warning in the face of a person mislabelling
the content. A DTD cannot tell if a writer has used a caution where a
warning is appropriate or vice versa. Similarly, a validator cannot
determine if a warning has been omitted where one is called for. It
can tell that a warning element is not present, but it has no way to
know that there should have been a warning in the first place.