## The Forum for Discussion about The Third Manifesto and Related Matters

Forum breadcrumbs - You are here:
Please or Register to create posts and topics.

# Which value?

Let me state the opposing positions. Number 1 from TTM and Database Explorations, and number 2 as my suggestion from various of these "Which....?" topics. :

1. Every scalar values is an ordered pair of Type name and Individual
2. All scalar values are Individuals

Which is OK, but I don't like the dependence on the word Scalar. So let me try this wordage:

1. The simplest a value can be (and have any meaning) is to be an ordered pair of a Type name and an Individual
2. The simplest a value can be (and have any meaning) is to be an Individual

Now, let me suggest that we are both wrong (well, technically I am right, but I won't brag ;-) ). I suggest that the real situation is the following:

0. The simplest a given value can be (and have any (useful) meaning) depends on the particular value in question— sometimes just an Individual is sufficient, sometimes (often in fact) a Type name and Individual is required, and sometimes more, a triple of  such as <Attribute Name, Type Name, Individual> say.

And, if that proposition is true, why stop at triples?. E.g. maybe, as David suggests, sometimes you need larger ordered sequences (such as ``17 >> age >> years >> range >> oldest >> person >> car driver >> not`for a Value to be truly meaningful.

Quote from Paul Vernon on December 6, 2021, 11:15 am

[...]

Which is OK, but I don't like the dependence on the word Scalar. So let me try this wordage:

1. The simplest a value can be (and have any meaning) is to be an ordered pair of a Type name and an Individual
2. The simplest a value can be (and have any meaning) is to be an Individual

Now, let me suggest that we are both wrong (well, technically I am right, but I won't brag ;-) ). I suggest that the real situation is the following:

0. The simplest a given value can be (and have any (useful) meaning) depends on the particular value in question— sometimes just an Individual is sufficient, sometimes (often in fact) a Type name and Individual is required, and sometimes more, a triple of  such as <Attribute Name, Type Name, Individual> say.

And, if that proposition is true, why stop at triples?. E.g. maybe, as David suggests, sometimes you need larger ordered sequences (such as ``17 >> age >> years >> range >> oldest >> person >> car driver >> not`for a Value to be truly meaningful.

In computational terms, the simplest a given value can be and have any useful meaning is some representation (e.g., a string of bits; I'm hesitant to use 'Individual' here as it's a term used by logicians not computer science -- or at least not the way logicians use it) paired with a (reference to a) type.

That's true even if the type isn't in the computer but is only in the programmer's mind, and it's true in the absence of an attribute name.  Terms in expressions, for example, have no associated attribute name but certainly have an identifiable type, though type inference mechanisms can also determine an associated "attribute name" from operator definitions (assuming parameters are named.)