Pointers in C can often be difficult to understand—I certainly had a learning curve and am continuing to learn. However, I had a thought that may help some by comparing a common experience and wanted to share.
A pointer in C behaves just like a word in any spoken language which refers to a physical object or multiple objects and the uniqueness of each object (e.g Skippy the dog, Mittens and Tiger the cats, fork number 5). The word itself does not contain the physical object and its uniqueness but only communicates the existence of the physical object and its uniqueness. The pointer itself does not contain the physical address and its value but only communicates the existence of the physical address and its value.
The better analogy is that people live in houses and houses have addresses, and I can use an address to find someone’s house.
Whether the pointed data by the pointer is valid or not is… not the point. In all languages I can think of, dereferencing an invalid pointer like a pointer to the wrong address per the type and alignment is never valid. Your analogy does not improve on historical analogies and it is wrong.
To be a bit more specific, a word is nothing more than a set of symbols (physical) which indicate an agreed value (abstract/reference). A pointer is the word and the agreed value is the reference of the object. The object the pointer points to is the existence of the real object (physical) and its value(s) which exists regardless of abstract references.
Pattern: physical -> reference -> physical -> value
Double pointer pattern: physical -> reference -> physical -> reference -> physical -> value
Etc…
A word’s meaning can change through time as cultures rise and fall, for the temporary purpose of encrypting conversations (e.g. the word “dog” can point to the agreed value of “fork” or “7”), or even misidentification.
That is a simple and good analogy, and, yes, perhaps better than my own.
Whether the pointed data by the pointer is valid or not is… not the point.
Nor is that the point of my analogy, but I do see how you inferred that point. Your criticism has helped me identify a flaw in how I express the analogy.
It’s not even an analogy; pointers and reference mechanics are the same concept in programming and linguistics. See the page on referents for an example blend of viewpoints.
Pronouns are pointers. “Let us (let’s) move it over there.” Both “us” and “it” indirectly refer to something else by a new name. Like pointers, the pointees are defined by some context external to that sentence/statement (usually earlier sentences/statements or some other actions). The meaning of “us” and “it” can change as well in different contexts, and as such, those words are not bound to one value (and “rebinding” those words by changing contexts does not change the values they were previously bound to).
Yes! 😁
One could even say that the pronouns “him” and “her” could indicate specific types of pointers, like int and char.
The pointer itself does not contain the physical address and its value but only communicates the existence of the physical address and its value.
So you’re telling me
0xDEADBEEFdoes not contain the physical address of the thing it points to?It’s a technicality about the pointer type. You can cast the type away which typically doesn’t change the actual value (but I’m pretty sure that causes undefined behavior)
For your example,
int x = 0xDEADBEEF;signifies the integer -559038737 (at least on x86.)char *p = (char*)0xDEADBEEF;on the other hand may or may not point to the real memory address 0xDEADBEEF, depending on factors like if the processor is using virtual or real addressing, etcYes and no, according to the analogy. Just as a word contains the idea of an object, but not the object itself.
I tried seeing if the analogy could be made to work but I am unaware of any natural language with an analog to pointers. Natural language references generally frame the concept that serves as a pointer in relation to the object rather than as a particular object.
Let’s make an agreement that the English word “coffee” now means 65. The reference of any word within any language only exists because we make an agreement with ourselves or with others for what a particular set of symbols gives reference to. Without an agreement, all the symbols we use mean nothing more than an arbitrary scratch on a table.
A pointer is the word, which means nothing by itself other than a grouping of letters exist, and the reference to an address is the reference to an object, which exists independently from the object—a reference to a fictional object can exist and a real object can exist without a reference.
Human language has 2 layers. Signifier and signified. Computers have 3. Signifier (variable name), memory address, (no human analog) and signified. (stored value) A pointer would be a signifier (word) for which the signified is the non-existent human analog to a memory address. You can have a word that refers to a thing that doesn’t exist (e.g. unicorn) but it won’t help the analogy.
In your example, the signifier ‘coffee’ is akin to a variable name. It is a signifier that dereferences to a value within the scope of our agreement. We can agree it holds the stored value of 65, allowing us to say something like butter = coffee - 60, bring me ‘butter’ eggs, but we don’t have a memory address system so it’s not the same process as a computer.
For a computer, there is a middle step to a variable dereference. It sees ‘coffee’ and first looks at the table of references, finds the variable ‘coffee’ has it’s value stored at address 1
['coffee':1, 'butter':88, 'unicorn':73]and then pulls the value (65) from the memory at address 1. A pointer would be another variable, such as ‘*monkey’ (e.g.*monkey =, which can be read as “pointer monkey shall hold the value of the address at which coffee is stored”) which would be added into the reference table as a variable stored at another address, say address 2, and would store in address 2 a value of (1).word address value coffee 1 65 *monkey 2 1Language just doesn’t have the analogous concept.
we don’t have a memory address system so it’s not the same process as a computer.
Just because we don’t have to consciously think about where data is stored in our brains doesn’t mean that our brains don’t have a memory address system. How often do you have to consciously maintain your heart beat, calculate the orientation of your body, or identify that the two objects in front of you are similar enough to be a set of objects—two dogs. You don’t. They are unconscious/automatic functions.
In a computer, checking a memory address would be closer to what we would call ‘conscious’ actions in a person. It is performed actively by software. In a brain, we still don’t really have a full understanding of how memory is stored, but we at least know it isn’t part of our experience that we have explicitly defined storage locations that we address. If it were, we’d have a name for it. Computers and brains are metaphorically similar but structurally and functionally quite different.
You have a very good point. I would personally say that there is a physical difference between a computer and human, but that they are logically equivalent, within the scope of our discussion, even if the human processes are unconscious, because, objectively, the processes are occurring and have logical equivalence in that they are both memory address systems; just as a number and letter can have logical equivalence as symbols.
Also, I realize that I do not have the evidence that our brains do operate on a memory address system, and since it is my claim the burden of proof is on me. However, lack of evidence on my part is not evidence for an opposing idea, meaning the correct answer, for now, is “I don’t know.”
Have a wonderful day! 😁
