At 18:21 13/05/04 +0100, Duncan Coutts wrote:
Something which wasn't mentioned but is quite useful is type-specialised xml parsers. ...
HaXML has something like this. I'd suggest that something like this wouldn't necessarily be *part* of an XML library, but an additional XML library that uses a generic XML framework. But maybe that's what you meant? As it happens, it's not part of the requirements I'm looking at, because my aim is extract information from XML to another internal format, but I acknowledge the possible value of this. Maybe it's worth reviewing what Wallace&Runciman say about this (section 3.4 of their paper [1]). I don't know if any of this work has been updated to take account of XML schemas. #g -- [1] http://www.cs.york.ac.uk/fp/HaXml/icfp99.html At 18:21 13/05/04 +0100, Duncan Coutts wrote:
Something which wasn't mentioned but is quite useful is type-specialised xml parsers. Some tools manipulate xml generically, giving you back some DOM tree which is great if you are writing a general purpose xml parser. However most uses know exactly what DTD/Schema/Type they are dealing with and would like to get their own data type back from the parser (as well as having the parser validate it). This allows you to use the parser/pretty printer in a similar way to ordinary read/show. Other people have pointed out that this should make xslt-style transformations really easy (and type safe). (Automatically deriving readXML/showXML would be nice!)
Some Haskell xml libs/tookits have tools for converting DTD<->Haskell types.
I suggest this would be a very useful feature of a standard xml library.
Duncan
------------ Graham Klyne For email: http://www.ninebynine.org/#Contact