On 14/02/2011 12:09, Vasu Chakkera wrote:
Dear David/Mike.
Thanks for this..
Yeas, It does kind of confuse the reader..
Yes. There was a healthy debate about it. I was opposed to allowing it.
It's a class of question we sometimes call "paternalism" - should one
add rules to the spec that disallow things whose meaning is perfectly
well-defined, but which could be perceived as bad practice? The WGs as a
whole sometimes incline one way on such questions and sometimes the
other - and I must admit that as an individual member of the WG, I'm not
entirely consistent myself.
Generally, orthogonality in language design is a good thing, which means
one tends to avoid arbitrary restrictions, which means one tends to
avoid paternalism.
So, for example, XPath decided that in axis steps, the axis and the
NodeTest should be orthogonal, which makes it legal to write
attribute::comment(). A paternalistic design would have banned the
combinations that can never select anything.
Michael Kay
Saxonica
--~------------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
To unsubscribe, go to: http://lists.mulberrytech.com/xsl-list/
or e-mail: <mailto:xsl-list-unsubscribe(_at_)lists(_dot_)mulberrytech(_dot_)com>
--~--