- From: David Carlisle <davidc@nag.co.uk>
- Date: Mon, 27 Nov 2006 11:45:52 GMT
- To: juanrgonzaleza@canonicalscience.com
- Cc: www-math@w3.org
> I am really confused. You cited several parts of the spec, but you didn't really say what parts of them you found confusing. ci and csymbol can be used in the same places, and take the same attributes and same content. As you said ci is intended to denote more general identifiers, and csymbol is intended to denote symbols with an external definition, but the mathematical distinction between these two is not precise, and so the usage of the two elements will display some personal (or tool) preferences, especially as MathML1 only had ci so some tools may have a preference for generating ci. In general if you have f(x) then you'd probably expect to use ci for f as a generic identifier, and if you have dx/dt (and didn't want to use <diff/> for some reason) you'd expect to use csymbol for diff. But it depends on circumstances, if the math fragment is in a book and says on page one, let f denote Ramanujan's F-function. http://mathworld.wolfram.com/Ramanujanf-Function.html ... then you may want to use csymbol for f Similarly if you are discussing various definitions of differential operator over different domains, you may want to use <ci>diff</ci> to refer to a generic reference to all of them (or at least one unspecified one) It's just a matter of degree, and you can't really say anything at all about the usage in a small fragment taken out of context. For both ci and csymbol, there is no requirement that the definitionURL refers to any machine readable definition, so the extent to which you can mechanically process content mathml depends on many more factors than whether ci or csymbol is used in the markup. This is an essential feature of any Mathematical encoding, not a weakness in MathML. If you try to be complete then provably you can't express anything interesting. David
Received on Monday, 27 November 2006 11:46:10 UTC