Cart
[Home] [By Thread] [By Date] [Recent Entries]
On Tue, 25 Aug 2009 22:39:51 +0100, Michael Kay wrote: >> (Mind you, I'm not suggesting that XML is free of "we must control" >> attitudes; see, for instance, W3C XML Schema, in which the >> collection of primitive types are all you get, unless you can >> convince the Schema WG to add *your* favorite unrelated >> primitive datatype to the collection). > > XML Schema 1.1 permits vendor-defined data types, and if the vendors choose > to provide mechanisms for defining them, user-defined data types. Saxon 9.2 > makes a small start with facilities that allow you to define your own > lexical representations of existing types, e.g. 1,234 for integers, on|off > for booleans, or 25/8/2009 for dates. Mea maxima culpa. My first draft included "1.0"; I rewrote the sentence for clarity, and lost specificity. I will ... refrain from going off on the W3C XML type collection. I have an argument on that score, but it needs a specification and an implementation, and my time is so far not my own that only the specification is completed, at present. > Of course, such things only become really useful when the mechanisms for > defining the extensions are standardised across products: but I'm a great > believer in the principle of providing extensibility first, and then > standardising the extensions (or preferably, the extensibility mechanisms) > that prove popular. It's like waiting to see where people want to walk > before you lay the footpaths. Agreed. I think this is a good analogy for how MathML and SVG were able to achieve such prominence as they have. My argument, in the message to which yours responds, was to the effect that I thought HTML ought to also have such an extensibility mechanism (if it were algorithmically transformable to XML namespaces, all the better). Amy!
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] |
