Recognizer v4 API calls the entity "parsed data" that is called "token" here. And it doesn't specify where these data pieces may be located.
Token is a tuple of data objects (of the known data types). In general, the data objects can be located not only in the data stack or in the floating-point stack, but also in the memory.
An excerpt from the 3.1 Data types section:
A data type identifies the set of permissible values for a data object. It is not a property of a particular storage location or position on a stack. Moving a data object shall not affect its type.
If a program provides all required actions to work with a token, nothing prevents a program to use other locations beyond the standard stacks. So, some components of the token can be located, for example, in some user-defined stack.
But it is not compatible with some methods to create a token descriptor (see #10). And also, THROW will not discard the components beyond the standard stacks.
So, the question is: should the location of token components be tied to the standard stacks only?
If it should, — we have more approaches to create descriptors and handling tokens.
At the moment I think, it should.
Or can anybody provide a useful use-case for the loosely case?
Recognizer v4 API calls the entity "parsed data" that is called "token" here. And it doesn't specify where these data pieces may be located.
Token is a tuple of data objects (of the known data types). In general, the data objects can be located not only in the data stack or in the floating-point stack, but also in the memory.
An excerpt from the 3.1 Data types section:
If a program provides all required actions to work with a token, nothing prevents a program to use other locations beyond the standard stacks. So, some components of the token can be located, for example, in some user-defined stack.
But it is not compatible with some methods to create a token descriptor (see #10). And also,
THROWwill not discard the components beyond the standard stacks.So, the question is: should the location of token components be tied to the standard stacks only?
If it should, — we have more approaches to create descriptors and handling tokens.
At the moment I think, it should.
Or can anybody provide a useful use-case for the loosely case?