[whatwg] 1. Re: New attributes would degrade better than new elements

[Francis Boumphrey]

I've been following this thread with a little confusion, because I'm not
sure what the argument is. However as one of the people involved in the
original XHTML spec. I do know what our intentions were so perhaps the
following may be useful.

HTML was meant to be extensile. We fully expected new elements to be added
in the future, and made provision for this. We also realized that we could
not predict future developments!

As far as possible, HTML was only to be used for semantic markup; all use
of elements or attributes for styling with the exception of the style
attribute (and we had a long discussion on this) were deprecated. We
decided, and this is formalized in HTML5, that the style attribute was too
valuable for development purposes, and too useful for single instance
styling.

While on this we did have a long discussion on the use of tables for page
layout, but decided to kick this 'can of worms' down the road. As far as I
can see it's still being kicked! (IMO it's too useful for creating 3 column
layout. If you look at the 'view source' W3C site home page, they use
multiple <div> tags for column creation, but notice that the borders are
not completley smooth, and the code gets quite complex)

CSS Grid and Columns properties address this, but they are still not  fully
implemented

All styling was to be taken care of by CSS and style sheets.

The <div> element was to be used as a general wrapper element. We fully
expected it to be refined and made more specific and granular in future
specs (as it has been in HTML5)

In keeping with the philosophy of both XML and SGML, element attributes
should define properties of elements not semantic content.

Our chief problem was that we were 'shackled' by having to use a formal XML
DTD for normative descriptions. HTML5 and the WhatWG have cut through this
Gordian Knot!

All this information is in the W3C archives, but I hope this summary may be
useful.

Frank