Hi albert,

>I grant you that the class definition syntax must be extended to provide for containership. My point is that an XML representation of the SCX/VCX is a cumbersome solution, while the extension of the class definition syntax is the best approach.

May you're right that a XML representation of the SCX tables is cumbersome, but there are more possible ways define classes in a XML document. I was more thinking about a readable hierarchy which can be easely translated to either PRG or VCX.

Walter,

>Nested class definitions could be supported with the following syntax:
>

>DEFINE CLASS ClassName1 AS ParentClass | DEFINE CLASS ... ENDDEFINE [OLEPUBLIC]
>  [[PROTECTED | HIDDEN PropertyName1, PropertyName2 ...]
>     [Object.]PropertyName = eExpression ...]
>  [ADD OBJECT [PROTECTED] ObjectName AS ClassName2 [NOINIT]
>     [WITH cPropertylist]]...
>  [[PROTECTED | HIDDEN] FUNCTION | PROCEDURE Name[_ACCESS | _ASSIGN]
>  | THIS_ACCESS [NODEFAULT]
>     cStatements
>  [ENDFUNC | ENDPROC]]...
>ENDDEFINE
>

>
>>Hi albert,
>>
>>According to the docs, this falls short when using containership. Oddly enough there are a few things you can't define with PRGs (containership) and a few things you can't define with VCX files (As I recall correctly some new options in VFP 7).
>
>>
>>The idea of it is just to build an easy bridge between the strategies. Well in a sense this is XML since it is ASCII and you can describe a table in it.
>>
>>Walter,
>>
>>>There is no need to use XML. There is already a standard text representation of a VFP class:
>>>
>>>

>>>DEFINE CLASS ClassName1 AS ParentClass [OLEPUBLIC]
>>>  [[PROTECTED | HIDDEN PropertyName1, PropertyName2 ...]
>>>     [Object.]PropertyName = eExpression ...]
>>>  [ADD OBJECT [PROTECTED] ObjectName AS ClassName2 [NOINIT]
>>>     [WITH cPropertylist]]...
>>>  [[PROTECTED | HIDDEN] FUNCTION | PROCEDURE Name[_ACCESS | _ASSIGN]
>>>  | THIS_ACCESS [NODEFAULT]
>>>     cStatements
>>>  [ENDFUNC | ENDPROC]]...
>>>ENDDEFINE
>>>

>>>
>>>
>>>>Hi gerry,
>>>>
>>>>>A library is a library, whether it's in the form of a .PRG or a .VCX.
>>>>>
>>>>>It's like asking a "user": "Would you like an SQL Server or an Oracle backend ?" Huh?
>>>>>
>>>>>Some people don't "want" to give up VCX's, etc. ... but I'm still waiting for a single concrete reason as to why they need to be kept around (other than "that's they way it's always been").
>>>>
>>>>Comming to think of the issue. Since, tables can be described with XML, and XML is sort of an ASCII format. Can XML fill the GAP between PRGs and VCX based classes ???
>>>>
>>>>Walter,