W3C home > Mailing lists > Public > www-dom@w3.org > October to December 1997

Re: Comments and requests.

From: Mike Champion <mcc@arbortext.com>
Date: Fri, 17 Oct 1997 14:10:11 -0400
Message-Id: <97Oct17.140909edt.18820@thicket.arbortext.com>
To: Alexandre Rafalovitch <arafalov@socs.uts.EDU.AU>, www-dom@w3.org
At 07:17 AM 10/17/97 -0400, Alexandre Rafalovitch wrote:
The DOM is built from the output of and XML or HTML parser.  Only "well
formed" XML or HTML may be meaningfully represented in the DOM.  The major
browsers, for good business reasons, I suppose, attempt to "make sense" out
of HTML that is not well formed, but don't count on the DOM doing anything
along these lines. 

Mike Champion

>On Thu, 16 Oct 1997, Paul Grosso wrote:
>> At 21:27 1997 10 15 -0400, Alexandr>> >These are comments on DOM Level 1
document as of 9th of October draft.
>> >
>> >  Am I correct in understanding that comment node should be generated for
>> >  that and returned in getChildren() call, but not in getAttributes()
>> >  call. The alternative is to not represent in-tag comment in DOM. What
>> >  about error nodes. (I know they should not happen, but....)
>> e Rafalovitch wrote:

>> I don't really know what you mean by error nodes, but in general it may
>> not be possible to define a DOM on erroneous input.  Certainly, the input
>> must be good enough to model before we can define a document object model
>> for it.
>What about the case when it is possible. For example, if you look at html
>source code with errors in Netscape, you would notice, it recovers from
>errors by ignoring part of the input (it is shown in different color).
>This way, you can see where was the error. In the same way an Error node
>would just contains text representation of skipped part and can be ignored 
>by the application processor if not needed. Most probably it could happen
>on any level.
>  Alex.
Received on Friday, 17 October 1997 14:09:45 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 20 October 2015 10:46:03 UTC