[
Lists Home |
Date Index |
Thread Index
]
At 02:10 PM 12/4/2002 -0500, W. E. Perry wrote:
>Well of course it interferes--specifically, it tightly constrains
>reuse--unless
>that reuse is itself predicated on throwing aside the datatypes provided,
>and isn't
>that petitio principii?
petitio principii = "begging the question" = assuming the conclusion in the
premise?
My question was:
"Why does the presence of a data type prevent reuse, since you can always
throw it away? Can you show me an example where this causes problems?"
The data type provides additional information about the data. You can feel
free to discard this information if you don't need it, or to use it to
reinterpret the data, or to use the data as is. I am trying to think of a
concrete example where the presence of this data really gets in the way.
I think Jeff's point was that it causes overhead, since the additional
information is just being thrown away. That's true. And in specifications
where data types are not particularly useful, it adds needless complexity.
Both of these are concrete points that I understand and accept.
But I don't yet understand how the presence of a datatype can make reuse
more difficult - except, perhaps, by requiring a cast. Can you help me with
a concrete example or two?
Jonathan
|
