>  Are you suggesting the first is ("a", "b", "c", undef)?  Yes, 

undef is a Perlism.

>  indeed, that _is_ different than ("a", "b", "c", "").  
>  However, the original data was:
>  
>      "a,b,c,"

The original data was ...Ã?121314Ã?121314Ã?121314Ã? where Ã? is the delimiter which is the s
ame as Jeff's a,b,c, that is <cr>  terminated, like a line of text being read.

>      
>  How do you express "undef" in a stringified representation of data?
>  Hint: only if you pre-define a means to do so, defining a 
>  semantic rule to the data _after_ it's been parsed.  
>  Otherwise, it _is_ an element that is just an empty string at 
>  the end.  Which is, exactly what Tcl gives you.

Well then I have to disagree because an empty string implies that something is/was there.  A
nalogous to writing a string then deleting it in a field, a zero-byte string remains.  But w
e now that is not the case.  However a better representation would be NULL because we know t
hat nothing is there except a carriage return.  So how do you express a NULL?  I think Tcl i
s presupposing that something is there.

Mike...
_______________________________________________
ActiveTcl mailing list
ActiveTcl@[...].com
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs