- From: David Carlisle <davidc@nag.co.uk>
- Date: Wed, 10 Nov 1999 19:18:32 GMT
- To: devlinh@nwu.edu
- CC: www-math@w3.org, strotman@nu.cs.fsu.edu
> Wouldn't this strip apply of some of its power to capture semantics? > Will <apply> become like a open paren and <apply/> like a close paren, > generic in the sense of serving as a container for either functional or > relation expressions? I don't think any semantics is lost. Especially if you view relations as boolean valued functions, the difference between a relation and a function is somewhat hard to detect. Even if you view relations as distinct from functions, the current restrictions mean that <= is a relation and so used with reln, but if you were to construct the operation `< or =' from `<' `=' and `or' then the result is not considered to be a relation any more and so would be applied with apply not reln. `relations' such as < usually get rather different rendering rules than binary operations such as +. but a MathMl renderer does not need the reln/apply distinction to detect this, as the renderer anyway needs to know any special layout requirements of each mathml content construction. David
Received on Wednesday, 10 November 1999 14:22:15 UTC