Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!ucsd!hub!eiffel!bertrand
From: bertrand@eiffel.UUCP (Bertrand Meyer)
Newsgroups: comp.lang.eiffel
Subject: Re: The indexing clause
Message-ID: <181@eiffel.UUCP>
Date: 15 Jul 89 18:31:46 GMT
References: <180@eiffel.UUCP> <120001@gore.com>
Organization: Interactive Software Engineering, Santa Barbara CA
Lines: 33

From <120001@gore.com> by jacob@gore.com (Jacob Gore):

> It is not clear to me that in a database [of reusable classes]
> lack of information [about a certain property of a class]
> must imply negation.  Why is it better to
> assume that the lack of a "representation:" [index entry] means
> "$NONE$" rather than "$UNKNOWN$" (or, perhaps,
> simply "$UNSPECIFIED$")?

I agree. In my hypothetical query language, a query pair of the form
<index, $NONE$> was meant as expressing absence of information
about ``index'', not negation. $ABSENT$ or $UNSPECIFIED$ would
be less ambiguous keywords than $NONE$.

True, one may argue that in the example at hand (the ``representation''
entry for a deferred class), what is desired is indeed negation,
not just absence:

    Not only does a deferred class have no ``representation'' property:
	not having a representation is one of its properties.

(It took me some time to phrase this right.) The indexing clause
is not needed, however, to capture such a negative property.
which is readily deduced from the deferred nature of the class.

Clearly, any browsing tool must have access to the information that
the class is deferred, and the query language should allow
a user to specify that the target of a query may, must or may not
be deferred.
-- 

-- Bertrand Meyer
bertrand@eiffel.com