[Bug 1621] no algebra of types

http://www.w3.org/Bugs/Public/show_bug.cgi?id=1621





------- Additional Comments From fred.zemke@oracle.com  2005-07-19 21:15 -------
This comment was on section 2.4.3, "Content models".

Subsequently, I have found 8.3.2 "Subtype and type equality",
which does consider the algebra of types.  The solution there, to use known 
algorithms for regular expressions and finite state automata, looks correct
and is actually algorithmic, unlike my proposal to specify the axioms from
which a type algebra could be deduced.  (How uncharacteristic of the 
Formal Semantics to actually suggest an algorithm to use to implement
it!)  I now suggest that this comment could be resolved by:

1. Putting a pointer from 2.4.3 to 8.3.2 so the reader knows where to find
a discussion of the algebra of types.

2. Perhaps a note or example in 8.3.2 that all the expected judgments about 
the type hierarcy, such as |- (Type1, empty) = Type1, can be deduced using
the regular expressions algorithms.

Received on Tuesday, 19 July 2005 21:15:46 UTC