RE: [xml-dev] Overloaded functions in XPath 1.0?

["Michael Kay" <michael.h.kay@ntlworld.com>]

If you mean a single function getValue() that sometimes returns a string and sometimes a number, you can do this in a weakly-typed language like XPath 1.0. If fact, the system-property() function does this. But I wouldn't advise it; it may give you problems when you move to XPath 2.0, which is more strongly typed, and it makes life more difficult for an optimizer.

 

This isn't really "Overloading" as the word is normally used.

 

Michael Kay

---

From: Jeremy Johnson [mailto:jeremy.johnson@imtc.gatech.edu]
Sent: 13 May 2004 15:32
To: xml-dev@lists.xml.org
Subject: [xml-dev] Overloaded functions in XPath 1.0?

An application I’m working on needs to be able to evaluate XPath 1.0 expressions.  In addition to the standard XPath core functions, we need to introduce an application-specific function that retrieves a value from a data structure internal to the app.  Lets call it ‘getvalue()’.  This value could be a String, Boolean or Number.  Would it be a violation of XPath practices for this ‘getvalue()’ function to be an overloaded function, returning String, Number and Boolean (e.g. three functions: string getvalue(), boolean getvalue(), number getvalue())?  My alternative would be to use unique function names such as getstring(), getboolean() and getnumber()...