>>>>> "Mike" == Michael Sperber <sperber@informatik.uni-tuebingen.de> writes:
>>>>> "Martin" == Martin Gasbichler <gasbichl@informatik.uni-tuebingen.de>
>>>>> writes:
>>>>> "Mike" == Michael Sperber <sperber@informatik.uni-tuebingen.de> writes:
>>>>> "Martin" == Martin Gasbichler <gasbichl@informatik.uni-tuebingen.de>
>>>>> writes:
>>>>> "Mike" == Michael Sperber <sperber@informatik.uni-tuebingen.de> writes:
Martin> It's heavily overloaded anyway, so why bother about one more?
Mike> Because adding worse to bad doesn't a good make :-)
Mike> More importantly, because it indirectly exposes the representation of
Mike> FILE-INFO even more than done through Olin's naming conventions for
Mike> records.
Martin> What are you talking about???
Mike> Overloading the existing procedures without providing explicit
Mike> FILE-INFO:... accessors exposes the fact that these properties aren't
Mike> fields of FILE-INFO; therefore we should provide them. If we do
Mike> provide them, what's the point of overloading the others?
Martin> The manual contains the record type of file-info, everybody can see
Martin> what it contains.
Mike> Yes, and that neither good nor necessary.
Maybe. The point is that we can't change it.
Martin> The mentioned procedures extract information from the fields
Martin> of the record. Adding them as additional fields would
Martin> introduce redundancy.
Mike> As would providing that functionality via overloading.
No. The procedures work on the existing data, they do not introduce
new data in addition to existing fields.
--
Martin
|