- From: TAN Kuan Hui <kuanhui@mobileworkspace.com>
- Date: Thu, 7 Feb 2002 11:32:53 +0800
- To: <www-ql@w3.org>
- Cc: "Jonathan Robie" <jonathan.robie@softwareag.com>, "Michael Rys" <mrys@microsoft.com>, "Per Bothner" <per@bothner.com>
> >At 02:02 PM 2/6/2002 -0800, Michael Rys wrote: >>Function libraries do not need set of variables to be passed. Function >>libraries provide functions that have parameters. >> agree and theoretically isn't a good practice for functions to rely on global variables for good reusuability. However, from the grammar, we can't pass in Query Prolog components like schema/namespace declaration as function parameters; quite the contrary, function definitions are always encapsulated within a Query Prolog, wouldn't that make functions less reusable say for a different set of schemas or namespaces ? I need time to further think on this. > >Here is one example: I wrote a function library for querying RDF. Almost >every function in that library needed to be able to reference a common set >of assertions. I had to pass the assertions into every function as a >parameter, but first I had to normalize them, so referring to the source >was not appropriate. > Hi Jonathan, you really lost me there, what do you mean exactly by passing assertions into functions as parameter ? We can only pass variable and at best a typed variable right ? Kuan Hui
Received on Wednesday, 6 February 2002 22:33:55 UTC