Re: Nearest Common Ancestor Report (XDb1's $1000 Challenge)

From: Neo <neo55592_at_hotmail.com>
Date: 18 May 2004 12:33:14 -0700
Message-ID: <4b45d3ad.0405181133.6dbab988_at_posting.google.com>


> > > you seem to desire the possibility to enter untyped data
> > All things in XDb1 are typed/classified.
> But I didn't use the word "classified", I used the word "untyped".

In XDb1's data model type, class and domain are the same thing. Even C.J. Date begins one of his chapters by saying "domain is essentially a data type".

> > > Remove types, and nothing prevents your user from entering "banana" as
> > > John's age.

Besides providing automatic checking for "basic" types, there is a fundamental difference between RM and XDb1's data model. RM sets rules on what data can be accepted. XDb1 accepts the data given to it and requires app logic to properly check/classify it.

Thus, in RM, if age has a data type of integer, it will accept 8 but not "eight".

In XDb1, the app logic would need to determine the class of the data being given, look at the class of existing ages and then make a decision. For example, given "eight", it would still can classify it as an age but the second classification would be word instead of integer.

In RM, data can have domain(s?) and data type(s?). In XDb1, data have class(es) and data type is simply an additional classification.

Even given the concept of data types in RM, nothing prevents the user from entering "banana" as a color (expect for app logic). Received on Tue May 18 2004 - 21:33:14 CEST

Original text of this message