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

RE: SD2 - Structured Attributes

From: Andrew Layman <andrewl@microsoft.com>
Date: Mon, 19 May 1997 12:12:32 -0700
Message-ID: <7BB61B44F197D011892800805FD4F7927DD996@RED-03-MSG.dns.microsoft.com>
To: "'Dave Hollander'" <dmh@hpsgml.fc.hp.com>, w3c-sgml-wg@w3.org
Thank you for the clarification on your question.  I will hopefully
address the whole topic of contents versus attributes at length later
today, so I'm going to confine myself here to describing what I meant,
and omit arguing in favor of or against it.

If an attribute has structure, its parts are parts of the attribute.
There is no inheritance implied, but rather containment.  It is the
attribute that is the attribute: its parts are simply parts (of the
attribute). 

My answer may sound peculiar, but that is because thinking about the
question in terms of inheritance is a completely wrong approach.  It is
as though I posed the question whether in this example below the
attribute "attr" inherits element-ness from the "thing" element.

	<thing attr="blue">

It doesn't inherit anything from "thing." It is an attribute of thing.
If it had structure, its parts would not inherit anything, they would
simply be parts.


> -----Original Message-----
> From:	Dave Hollander [SMTP:dmh@hpsgml.fc.hp.com]
> Sent:	Monday, May 19, 1997 8:06 AM
> To:	w3c-sgml-wg@w3.org
> Subject:	Re: SD2 - Structured Attributes 
> 
> 
> I sure can try.
> 
> Assuming alternative syntax proposals will be offered, I belive the
> issue 
> is a detail that should be added to the problem statement. 
> 
> The syntax that differentiates between contents and attributes 
> -should/should not- (choose 1) be inherited by children elements.
> 
> eg are these equivalent?
> 
> <head>
> <meta xml-att="true" contents="some contents"\>
> <meta xml-att="true" contents="some more contents"\>
> <meta xml-att="true" contents="some end contents"\>
> </head>
> 
> <head xml-att="true">
> <meta contents="some contents"\>
> <meta contents="some more contents"\>
> <meta contents="some end contents"\>
> </head>
> 
> 
> or
> 
> <*head>
> <meta contents="some contents"\>
> <meta contents="some more contents"\>
> <meta contents="some end contents"\>
> </head>
> 
> <head>
> <*meta contents="some contents"\>
> <*meta contents="some more contents"\>
> <*meta contents="some end contents"\>
> </head>
> 
> 
> Dave
> 
> > Could you expand on this point?  I don't understand it, and want to
> make
> > sure that I do.  Thanks.
> > 
> > > -----Original Message-----
> > > From:	Dave Hollander [SMTP:dmh@hpsgml.fc.hp.com]
> > > Sent:	Friday, May 16, 1997 12:19 PM
> > > To:	w3c-sgml-wg@w3.org
> > > Subject:	Re: SD2 - Structured Attributes 
> > > 
> > > 
> > > I we are to provide a distinction for attributes and content,
> should
> > > we have the ability to make that distinction for containers?
> > > This would make the HTML head element easy to express and process.
> > > I think Lee's post implied such an semantic.
> > > 
> > > Dave
> > 
Received on Monday, 19 May 1997 15:49:17 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:26 UTC